We’ve been having conversations for thousands of years. Whether to convey information, conduct transactions, or simply to check in on one another, people have yammered away, chattering and gesticulating, through spoken conversation for countless generations. Only in the last few millennia have we begun to commit our conversations to writing, and only in the last few decades have we begun to outsource them to the computer, a machine that shows much more affinity for written correspondence than for the slangy vagaries of spoken language.

Computers have trouble because between spoken and written language, speech is more primordial. To have successful conversations with us, machines must grapple with the messiness of human speech: the disfluencies and pauses, the gestures and body language, and the variations in word choice and spoken dialect that can stymie even the most carefully crafted human-computer interaction. In the human-to-human scenario, spoken language also has the privilege of face-to-face contact, where we can readily interpret nonverbal social cues.

In contrast, written language immediately concretizes as we commit it to record and retains usages long after they become obsolete in spoken communication (the salutation “To whom it may concern,” for example), generating its own fossil record of outdated terms and phrases. Because it tends to be more consistent, polished, and formal, written text is fundamentally much easier for machines to parse and understand.

Spoken language has no such luxury. Besides the nonverbal cues that decorate conversations with emphasis and emotional context, there are also verbal cues and vocal behaviors that modulate conversation in nuanced ways: how something is said, not what. Whether rapid-fire, low-pitched, or high-decibel, whether sarcastic, stilted, or sighing, our spoken language conveys much more than the written word could ever muster. So when it comes to voice interfaces—the machines we conduct spoken conversations with—we face exciting challenges as designers and content strategists.

Voice Interactions

We interact with voice interfaces for a variety of reasons, but according to Michael McTear, Zoraida Callejas, and David Griol in The Conversational Interface, those motivations by and large mirror the reasons we initiate conversations with other people, too (). Generally, we start up a conversation because:

• we need something done (such as a transaction),
• we want to know something (information of some sort), or
• we are social beings and want someone to talk to (conversation for conversation’s sake).

These three categories—which I call transactional, informational, and prosocial—also characterize essentially every voice interaction: a single conversation from beginning to end that realizes some outcome for the user, starting with the voice interface’s first greeting and ending with the user exiting the interface. Note here that a conversation in our human sense—a chat between people that leads to some result and lasts an arbitrary length of time—could encompass multiple transactional, informational, and prosocial voice interactions in succession. In other words, a voice interaction is a conversation, but a conversation is not necessarily a single voice interaction.

Purely prosocial conversations are more gimmicky than captivating in most voice interfaces, because machines don’t yet have the capacity to really want to know how we’re doing and to do the sort of glad-handing humans crave. There’s also ongoing debate as to whether users actually prefer the sort of organic human conversation that begins with a prosocial voice interaction and shifts seamlessly into other types. In fact, in Voice User Interface Design, Michael Cohen, James Giangola, and Jennifer Balogh recommend sticking to users’ expectations by mimicking how they interact with other voice interfaces rather than trying too hard to be human—potentially alienating them in the process ().

That leaves two genres of conversations we can have with one another that a voice interface can easily have with us, too: a transactional voice interaction realizing some outcome (“buy iced tea”) and an informational voice interaction teaching us something new (“discuss a musical”).

Transactional voice interactions

Unless you’re tapping buttons on a food delivery app, you’re generally having a conversation—and therefore a voice interaction—when you order a Hawaiian pizza with extra pineapple. Even when we walk up to the counter and place an order, the conversation quickly pivots from an initial smattering of neighborly small talk to the real mission at hand: ordering a pizza (generously topped with pineapple, as it should be).

Alison: Hey, how’s it going?

Burhan: Hi, welcome to Crust Deluxe! It’s cold out there. How can I help you?

Alison: Can I get a Hawaiian pizza with extra pineapple?

Burhan: Sure, what size?

Alison: Large.

Burhan: Anything else?

Alison: No thanks, that’s it.

Burhan: Something to drink?

Alison: I’ll have a bottle of Coke.

Burhan: You got it. That’ll be $13.55 and about fifteen minutes.

Each progressive disclosure in this transactional conversation reveals more and more of the desired outcome of the transaction: a service rendered or a product delivered. Transactional conversations have certain key traits: they’re direct, to the point, and economical. They quickly dispense with pleasantries.

Informational voice interactions

Meanwhile, some conversations are primarily about obtaining information. Though Alison might visit Crust Deluxe with the sole purpose of placing an order, she might not actually want to walk out with a pizza at all. She might be just as interested in whether they serve halal or kosher dishes, gluten-free options, or something else. Here, though we again have a prosocial mini-conversation at the beginning to establish politeness, we’re after much more.

Alison: Hey, how’s it going?

Burhan: Hi, welcome to Crust Deluxe! It’s cold out there. How can I help you?

Alison: Can I ask a few questions?

Burhan: Of course! Go right ahead.

Alison: Do you have any halal options on the menu?

Burhan: Absolutely! We can make any pie halal by request. We also have lots of vegetarian, ovo-lacto, and vegan options. Are you thinking about any other dietary restrictions?

Alison: What about gluten-free pizzas?

Burhan: We can definitely do a gluten-free crust for you, no problem, for both our deep-dish and thin-crust pizzas. Anything else I can answer for you?

Alison: That’s it for now. Good to know. Thanks!

Burhan: Anytime, come back soon!

This is a very different dialogue. Here, the goal is to get a certain set of facts. Informational conversations are investigative quests for the truth—research expeditions to gather data, news, or facts. Voice interactions that are informational might be more long-winded than transactional conversations by necessity. Responses tend to be lengthier, more informative, and carefully communicated so the customer understands the key takeaways.

Voice Interfaces

At their core, voice interfaces employ speech to support users in reaching their goals. But simply because an interface has a voice component doesn’t mean that every user interaction with it is mediated through voice. Because multimodal voice interfaces can lean on visual components like screens as crutches, we’re most concerned in this book with pure voice interfaces, which depend entirely on spoken conversation, lack any visual component whatsoever, and are therefore much more nuanced and challenging to tackle.

Though voice interfaces have long been integral to the imagined future of humanity in science fiction, only recently have those lofty visions become fully realized in genuine voice interfaces.

Interactive voice response (IVR) systems

Though written conversational interfaces have been fixtures of computing for many decades, voice interfaces first emerged in the early 1990s with text-to-speech (TTS) dictation programs that recited written text aloud, as well as speech-enabled in-car systems that gave directions to a user-provided address. With the advent of interactive voice response (IVR) systems, intended as an alternative to overburdened customer service representatives, we became acquainted with the first true voice interfaces that engaged in authentic conversation.

IVR systems allowed organizations to reduce their reliance on call centers but soon became notorious for their clunkiness. Commonplace in the corporate world, these systems were primarily designed as metaphorical switchboards to guide customers to a real phone agent (“Say Reservations to book a flight or check an itinerary”); chances are you will enter a conversation with one when you call an airline or hotel conglomerate. Despite their functional issues and users’ frustration with their inability to speak to an actual human right away, IVR systems proliferated in the early 1990s across a variety of industries (, PDF).

While IVR systems are great for highly repetitive, monotonous conversations that generally don’t veer from a single format, they have a reputation for less scintillating conversation than we’re used to in real life (or even in science fiction).

Screen readers

Parallel to the evolution of IVR systems was the invention of the screen reader, a tool that transcribes visual content into synthesized speech. For Blind or visually impaired website users, it’s the predominant method of interacting with text, multimedia, or form elements. Screen readers represent perhaps the closest equivalent we have today to an out-of-the-box implementation of content delivered through voice.

Among the first screen readers known by that moniker was the Screen Reader for the BBC Micro and NEEC Portable developed by the Research Centre for the Education of the Visually Handicapped (RCEVH) at the University of Birmingham in 1986 (). That same year, Jim Thatcher created the first IBM Screen Reader for text-based computers, later recreated for computers with graphical user interfaces (GUIs) ().

With the rapid growth of the web in the 1990s, the demand for accessible tools for websites exploded. Thanks to the introduction of semantic HTML and especially ARIA roles beginning in 2008, screen readers started facilitating speedy interactions with web pages that ostensibly allow disabled users to traverse the page as an aural and temporal space rather than a visual and physical one. In other words, screen readers for the web “provide mechanisms that translate visual design constructs—proximity, proportion, etc.—into useful information,” writes Aaron Gustafson in A List Apart. “At least they do when documents are authored thoughtfully” ().

Though deeply instructive for voice interface designers, there’s one significant problem with screen readers: they’re difficult to use and unremittingly verbose. The visual structures of websites and web navigation don’t translate well to screen readers, sometimes resulting in unwieldy pronouncements that name every manipulable HTML element and announce every formatting change. For many screen reader users, working with web-based interfaces exacts a cognitive toll.

In Wired, accessibility advocate and voice engineer Chris Maury considers why the screen reader experience is ill-suited to users relying on voice:

From the beginning, I hated the way that Screen Readers work. Why are they designed the way they are? It makes no sense to present information visually and then, and only then, translate that into audio. All of the time and energy that goes into creating the perfect user experience for an app is wasted, or even worse, adversely impacting the experience for blind users. ()

In many cases, well-designed voice interfaces can speed users to their destination better than long-winded screen reader monologues. After all, visual interface users have the benefit of darting around the viewport freely to find information, ignoring areas irrelevant to them. Blind users, meanwhile, are obligated to listen to every utterance synthesized into speech and therefore prize brevity and efficiency. Disabled users who have long had no choice but to employ clunky screen readers may find that voice interfaces, particularly more modern voice assistants, offer a more streamlined experience.

Voice assistants

When we think of voice assistants (the subset of voice interfaces now commonplace in living rooms, smart homes, and offices), many of us immediately picture HAL from 2001: A Space Odyssey or hear Majel Barrett’s voice as the omniscient computer in Star Trek. Voice assistants are akin to personal concierges that can answer questions, schedule appointments, conduct searches, and perform other common day-to-day tasks. And they’re rapidly gaining more attention from accessibility advocates for their assistive potential.

Before the earliest IVR systems found success in the enterprise, Apple published a demonstration video in 1987 depicting the Knowledge Navigator, a voice assistant that could transcribe spoken words and recognize human speech to a great degree of accuracy. Then, in 2001, Tim Berners-Lee and others formulated their vision for a Semantic Web “agent” that would perform typical errands like “checking calendars, making appointments, and finding locations” (, behind paywall). It wasn’t until 2011 that Apple’s Siri finally entered the picture, making voice assistants a tangible reality for consumers.

Thanks to the plethora of voice assistants available today, there is considerable variation in how programmable and customizable certain voice assistants are over others (Fig 1.1). At one extreme, everything except vendor-provided features is locked down; for example, at the time of their release, the core functionality of Apple’s Siri and Microsoft’s Cortana couldn’t be extended beyond their existing capabilities. Even today, it isn’t possible to program Siri to perform arbitrary functions, because there’s no means by which developers can interact with Siri at a low level, apart from predefined categories of tasks like sending messages, hailing rideshares, making restaurant reservations, and certain others.

At the opposite end of the spectrum, voice assistants like Amazon Alexa and Google Home offer a core foundation on which developers can build custom voice interfaces. For this reason, programmable voice assistants that lend themselves to customization and extensibility are becoming increasingly popular for developers who feel stifled by the limitations of Siri and Cortana. Amazon offers the Alexa Skills Kit, a developer framework for building custom voice interfaces for Amazon Alexa, while Google Home offers the ability to program arbitrary Google Assistant skills. Today, users can choose from among thousands of custom-built skills within both the Amazon Alexa and Google Assistant ecosystems.

As corporations like Amazon, Apple, Microsoft, and Google continue to stake their territory, they’re also selling and open-sourcing an unprecedented array of tools and frameworks for designers and developers that aim to make building voice interfaces as easy as possible, even without code.

Often by necessity, voice assistants like Amazon Alexa tend to be monochannel—they’re tightly coupled to a device and can’t be accessed on a computer or smartphone instead. By contrast, many development platforms like Google’s Dialogflow have introduced omnichannel capabilities so users can build a single conversational interface that then manifests as a voice interface, textual chatbot, and IVR system upon deployment. I don’t prescribe any specific implementation approaches in this design-focused book, but in Chapter 4 we’ll get into some of the implications these variables might have on the way you build out your design artifacts.

Voice Content

Simply put, voice content is content delivered through voice. To preserve what makes human conversation so compelling in the first place, voice content needs to be free-flowing and organic, contextless and concise—everything written content isn’t.

Our world is replete with voice content in various forms: screen readers reciting website content, voice assistants rattling off a weather forecast, and automated phone hotline responses governed by IVR systems. In this book, we’re most concerned with content delivered auditorily—not as an option, but as a necessity.

For many of us, our first foray into informational voice interfaces will be to deliver content to users. There’s only one problem: any content we already have isn’t in any way ready for this new habitat. So how do we make the content trapped on our websites more conversational? And how do we write new copy that lends itself to voice interactions?

Lately, we’ve begun slicing and dicing our content in unprecedented ways. Websites are, in many respects, colossal vaults of what I call macrocontent: lengthy prose that can extend for infinitely scrollable miles in a browser window, like microfilm viewers of newspaper archives. Back in 2002, well before the present-day ubiquity of voice assistants, technologist Anil Dash defined microcontent as permalinked pieces of content that stay legible regardless of environment, such as email or text messages:

A day’s weather forcast [sic], the arrival and departure times for an airplane flight, an abstract from a long publication, or a single instant message can all be examples of microcontent. ()

I’d update Dash’s definition of microcontent to include all examples of bite-sized content that go well beyond written communiqués. After all, today we encounter microcontent in interfaces where a small snippet of copy is displayed alone, unmoored from the browser, like a textbot confirmation of a restaurant reservation. Microcontent offers the best opportunity to gauge how your content can be stretched to the very edges of its capabilities, informing delivery channels both established and novel.

As microcontent, voice content is unique because it’s an example of how content is experienced in time rather than in space. We can glance at a digital sign underground for an instant and know when the next train is arriving, but voice interfaces hold our attention captive for periods of time that we can’t easily escape or skip, something screen reader users are all too familiar with.

Because microcontent is fundamentally made up of isolated blobs with no relation to the channels where they’ll eventually end up, we need to ensure that our microcontent truly performs well as voice content—and that means focusing on the two most important traits of robust voice content: voice content legibility and voice content discoverability.

Fundamentally, the legibility and discoverability of our voice content both have to do with how voice content manifests in perceived time and space.

In the 1950s, many in the elite running community had begun to believe it wasn’t possible to run a mile in less than four minutes. Runners had been attempting it since the late 19th century and were beginning to draw the conclusion that the human body simply wasn’t built for the task. 

But on May 6, 1956, Roger Bannister took everyone by surprise. It was a cold, wet day in Oxford, England—conditions no one expected to lend themselves to record-setting—and yet Bannister did just that, running a mile in 3:59.4 and becoming the first person in the record books to run a mile in under four minutes. 

This shift in the benchmark had profound effects; the world now knew that the four-minute mile was possible. Bannister’s record lasted only forty-six days, when it was snatched away by Australian runner John Landy. Then a year later, three runners all beat the four-minute barrier together in the same race. Since then, over 1,400 runners have officially run a mile in under four minutes; the current record is 3:43.13, held by Moroccan athlete Hicham El Guerrouj.

We achieve far more when we believe that something is possible, and we will believe it’s possible only when we see someone else has already done it—and as with human running speed, so it is with what we believe are the hard limits for how a website needs to perform.

Establishing standards for a sustainable web

In most major industries, the key metrics of environmental performance are fairly well established, such as miles per gallon for cars or energy per square meter for homes. The tools and methods for calculating those metrics are standardized as well, which keeps everyone on the same page when doing environmental assessments. In the world of websites and apps, however, we aren’t held to any particular environmental standards, and only recently have gained the tools and methods we need to even make an environmental assessment.

The primary goal in sustainable web design is to reduce carbon emissions. However, it’s almost impossible to actually measure the amount of CO₂ produced by a web product. We can’t measure the fumes coming out of the exhaust pipes on our laptops. The emissions of our websites are far away, out of sight and out of mind, coming out of power stations burning coal and gas. We have no way to trace the electrons from a website or app back to the power station where the electricity is being generated and actually know the exact amount of greenhouse gas produced. So what do we do? 

If we can’t measure the actual carbon emissions, then we need to find what we can measure. The primary factors that could be used as indicators of carbon emissions are:

1. Data transfer 
2. Carbon intensity of electricity

Let’s take a look at how we can use these metrics to quantify the energy consumption, and in turn the carbon footprint, of the websites and web apps we create.

Data transfer

Most researchers use kilowatt-hours per gigabyte (kWh/GB) as a metric of energy efficiency when measuring the amount of data transferred over the internet when a website or application is used. This provides a great reference point for energy consumption and carbon emissions. As a rule of thumb, the more data transferred, the more energy used in the data center, telecoms networks, and end user devices.

For web pages, data transfer for a single visit can be most easily estimated by measuring the page weight, meaning the transfer size of the page in kilobytes the first time someone visits the page. It’s fairly easy to measure using the developer tools in any modern web browser. Often your web hosting account will include statistics for the total data transfer of any web application (Fig 2.1).

The nice thing about page weight as a metric is that it allows us to compare the efficiency of web pages on a level playing field without confusing the issue with constantly changing traffic volumes. 

Reducing page weight requires a large scope. By early 2020, the median page weight was 1.97 MB for setups the HTTP Archive classifies as “desktop” and 1.77 MB for “mobile,” with desktop increasing 36 percent since January 2016 and mobile page weights nearly doubling in the same period (Fig 2.2). Roughly half of this data transfer is image files, making images the single biggest source of carbon emissions on the average website. 

History clearly shows us that our web pages can be smaller, if only we set our minds to it. While most technologies become ever more energy efficient, including the underlying technology of the web such as data centers and transmission networks, websites themselves are a technology that becomes less efficient as time goes on.

You might be familiar with the concept of performance budgeting as a way of focusing a project team on creating faster user experiences. For example, we might specify that the website must load in a maximum of one second on a broadband connection and three seconds on a 3G connection. Much like speed limits while driving, performance budgets are upper limits rather than vague suggestions, so the goal should always be to come in under budget.

Designing for fast performance does often lead to reduced data transfer and emissions, but it isn’t always the case. Web performance is often more about the subjective perception of load times than it is about the true efficiency of the underlying system, whereas page weight and transfer size are more objective measures and more reliable benchmarks for sustainable web design. 

We can set a page weight budget in reference to a benchmark of industry averages, using data from sources like HTTP Archive. We can also benchmark page weight against competitors or the old version of the website we’re replacing. For example, we might set a maximum page weight budget as equal to our most efficient competitor, or we could set the benchmark lower to guarantee we are best in class. 

If we want to take it to the next level, then we could also start looking at the transfer size of our web pages for repeat visitors. Although page weight for the first time someone visits is the easiest thing to measure, and easy to compare on a like-for-like basis, we can learn even more if we start looking at transfer size in other scenarios too. For example, visitors who load the same page multiple times will likely have a high percentage of the files cached in their browser, meaning they don’t need to transfer all of the files on subsequent visits. Likewise, a visitor who navigates to new pages on the same website will likely not need to load the full page each time, as some global assets from areas like the header and footer may already be cached in their browser. Measuring transfer size at this next level of detail can help us learn even more about how we can optimize efficiency for users who regularly visit our pages, and enable us to set page weight budgets for additional scenarios beyond the first visit.

Page weight budgets are easy to track throughout a design and development process. Although they don’t actually tell us carbon emission and energy consumption analytics directly, they give us a clear indication of efficiency relative to other websites. And as transfer size is an effective analog for energy consumption, we can actually use it to estimate energy consumption too.

In summary, reduced data transfer translates to energy efficiency, a key factor to reducing carbon emissions of web products. The more efficient our products, the less electricity they use, and the less fossil fuels need to be burned to produce the electricity to power them. But as we’ll see next, since all web products demand some power, it’s important to consider the source of that electricity, too.

Carbon intensity of electricity

Regardless of energy efficiency, the level of pollution caused by digital products depends on the carbon intensity of the energy being used to power them. Carbon intensity is a term used to define the grams of CO₂ produced for every kilowatt-hour of electricity (gCO₂/kWh). This varies widely, with renewable energy sources and nuclear having an extremely low carbon intensity of less than 10 gCO₂/kWh (even when factoring in their construction); whereas fossil fuels have very high carbon intensity of approximately 200–400 gCO₂/kWh. 

Most electricity comes from national or state grids, where energy from a variety of different sources is mixed together with varying levels of carbon intensity. The distributed nature of the internet means that a single user of a website or app might be using energy from multiple different grids simultaneously; a website user in Paris uses electricity from the French national grid to power their home internet and devices, but the website’s data center could be in Dallas, USA, pulling electricity from the Texas grid, while the telecoms networks use energy from everywhere between Dallas and Paris.

We don’t have control over the full energy supply of web services, but we do have some control over where we host our projects. With a data center using a significant proportion of the energy of any website, locating the data center in an area with low carbon energy will tangibly reduce its carbon emissions. Danish startup Tomorrow reports and maps this user-contributed data, and a glance at their map shows how, for example, choosing a data center in France will have significantly lower carbon emissions than a data center in the Netherlands (Fig 2.3).

That said, we don’t want to locate our servers too far away from our users; it takes energy to transmit data through the telecom’s networks, and the further the data travels, the more energy is consumed. Just like food miles, we can think of the distance from the data center to the website’s core user base as “megabyte miles”—and we want it to be as small as possible.

Using the distance itself as a benchmark, we can use website analytics to identify the country, state, or even city where our core user group is located and measure the distance from that location to the data center used by our hosting company. This will be a somewhat fuzzy metric as we don’t know the precise center of mass of our users or the exact location of a data center, but we can at least get a rough idea. 

For example, if a website is hosted in London but the primary user base is on the West Coast of the USA, then we could look up the distance from London to San Francisco, which is 5,300 miles. That’s a long way! We can see that hosting it somewhere in North America, ideally on the West Coast, would significantly reduce the distance and thus the energy used to transmit the data. In addition, locating our servers closer to our visitors helps reduce latency and delivers better user experience, so it’s a win-win.

Converting it back to carbon emissions

If we combine carbon intensity with a calculation for energy consumption, we can calculate the carbon emissions of our websites and apps. A tool my team created does this by measuring the data transfer over the wire when loading a web page, calculating the amount of electricity associated, and then converting that into a figure for CO₂ (Fig 2.4). It also factors in whether or not the web hosting is powered by renewable energy.

If you want to take it to the next level and tailor the data more accurately to the unique aspects of your project, the Energy and Emissions Worksheet accompanying this book shows you how.

With the ability to calculate carbon emissions for our projects, we could actually take a page weight budget one step further and set carbon budgets as well. CO₂ is not a metric commonly used in web projects; we’re more familiar with kilobytes and megabytes, and can fairly easily look at design options and files to assess how big they are. Translating that into carbon adds a layer of abstraction that isn’t as intuitive—but carbon budgets do focus our minds on the primary thing we’re trying to reduce, and support the core objective of sustainable web design: reducing carbon emissions.

Browser Energy

Data transfer might be the simplest and most complete analog for energy consumption in our digital projects, but by giving us one number to represent the energy used in the data center, the telecoms networks, and the end user’s devices, it can’t offer us insights into the efficiency in any specific part of the system.

One part of the system we can look at in more detail is the energy used by end users’ devices. As front-end web technologies become more advanced, the computational load is increasingly moving from the data center to users’ devices, whether they be phones, tablets, laptops, desktops, or even smart TVs. Modern web browsers allow us to implement more complex styling and animation on the fly using CSS and JavaScript. Furthermore, JavaScript libraries such as Angular and React allow us to create applications where the “thinking” work is done partly or entirely in the browser. 

All of these advances are exciting and open up new possibilities for what the web can do to serve society and create positive experiences. However, more computation in the user’s web browser means more energy used by their devices. This has implications not just environmentally, but also for user experience and inclusivity. Applications that put a heavy processing load on the user’s device can inadvertently exclude users with older, slower devices and cause batteries on phones and laptops to drain faster. Furthermore, if we build web applications that require the user to have up-to-date, powerful devices, people throw away old devices much more frequently. This isn’t just bad for the environment, but it puts a disproportionate financial burden on the poorest in society.

In part because the tools are limited, and partly because there are so many different models of devices, it’s difficult to measure website energy consumption on end users’ devices. One tool we do currently have is the Energy Impact monitor inside the developer console of the Safari browser (Fig 2.5).

You know when you load a website and your computer’s cooling fans start spinning so frantically you think it might actually take off? That’s essentially what this tool is measuring. 

It shows us the percentage of CPU used and the duration of CPU usage when loading the web page, and uses these figures to generate an energy impact rating. It doesn’t give us precise data for the amount of electricity used in kilowatts, but the information it does provide can be used to benchmark how efficiently your websites use energy and set targets for improvement.

Antiracist economist Kim Crayton says that “intention without strategy is chaos.” We’ve discussed how our biases, assumptions, and inattention toward marginalized and vulnerable groups lead to dangerous and unethical tech—but what, specifically, do we need to do to fix it? The intention to make our tech safer is not enough; we need a strategy.

This chapter will equip you with that plan of action. It covers how to integrate safety principles into your design work in order to create tech that’s safe, how to convince your stakeholders that this work is necessary, and how to respond to the critique that what we actually need is more diversity. (Spoiler: we do, but diversity alone is not the antidote to fixing unethical, unsafe tech.)

The process for inclusive safety

When you are designing for safety, your goals are to:

• identify ways your product can be used for abuse,
• design ways to prevent the abuse, and
• provide support for vulnerable users to reclaim power and control.

The Process for Inclusive Safety is a tool to help you reach those goals (Fig 5.1). It’s a methodology I created in 2018 to capture the various techniques I was using when designing products with safety in mind. Whether you are creating an entirely new product or adding to an existing feature, the Process can help you make your product safe and inclusive. The Process includes five general areas of action:

• Conducting research
• Creating archetypes
• Brainstorming problems
• Designing solutions
• Testing for safety

The Process is meant to be flexible—it won’t make sense for teams to implement every step in some situations. Use the parts that are relevant to your unique work and context; this is meant to be something you can insert into your existing design practice.

And once you use it, if you have an idea for making it better or simply want to provide context of how it helped your team, please get in touch with me. It’s a living document that I hope will continue to be a useful and realistic tool that technologists can use in their day-to-day work.

If you’re working on a product specifically for a vulnerable group or survivors of some form of trauma, such as an app for survivors of domestic violence, sexual assault, or drug addiction, be sure to read Chapter 7, which covers that situation explicitly and should be handled a bit differently. The guidelines here are for prioritizing safety when designing a more general product that will have a wide user base (which, we already know from statistics, will include certain groups that should be protected from harm). Chapter 7 is focused on products that are specifically for vulnerable groups and people who have experienced trauma.

Step 1: Conduct research

Design research should include a broad analysis of how your tech might be weaponized for abuse as well as specific insights into the experiences of survivors and perpetrators of that type of abuse. At this stage, you and your team will investigate issues of interpersonal harm and abuse, and explore any other safety, security, or inclusivity issues that might be a concern for your product or service, like data security, racist algorithms, and harassment.

Broad research

Your project should begin with broad, general research into similar products and issues around safety and ethical concerns that have already been reported. For example, a team building a smart home device would do well to understand the multitude of ways that existing smart home devices have been used as tools of abuse. If your product will involve AI, seek to understand the potentials for racism and other issues that have been reported in existing AI products. Nearly all types of technology have some kind of potential or actual harm that’s been reported on in the news or written about by academics. Google Scholar is a useful tool for finding these studies.

Specific research: Survivors

When possible and appropriate, include direct research (surveys and interviews) with people who are experts in the forms of harm you have uncovered. Ideally, you’ll want to interview advocates working in the space of your research first so that you have a more solid understanding of the topic and are better equipped to not retraumatize survivors. If you’ve uncovered possible domestic violence issues, for example, the experts you’ll want to speak with are survivors themselves, as well as workers at domestic violence hotlines, shelters, other related nonprofits, and lawyers.

Especially when interviewing survivors of any kind of trauma, it is important to pay people for their knowledge and lived experiences. Don’t ask survivors to share their trauma for free, as this is exploitative. While some survivors may not want to be paid, you should always make the offer in the initial ask. An alternative to payment is to donate to an organization working against the type of violence that the interviewee experienced. We’ll talk more about how to appropriately interview survivors in Chapter 6.

Specific research: Abusers

It’s unlikely that teams aiming to design for safety will be able to interview self-proclaimed abusers or people who have broken laws around things like hacking. Don’t make this a goal; rather, try to get at this angle in your general research. Aim to understand how abusers or bad actors weaponize technology to use against others, how they cover their tracks, and how they explain or rationalize the abuse.

Step 2: Create archetypes

Once you’ve finished conducting your research, use your insights to create abuser and survivor archetypes. Archetypes are not personas, as they’re not based on real people that you interviewed and surveyed. Instead, they’re based on your research into likely safety issues, much like when we design for accessibility: we don’t need to have found a group of blind or low-vision users in our interview pool to create a design that’s inclusive of them. Instead, we base those designs on existing research into what this group needs. Personas typically represent real users and include many details, while archetypes are broader and can be more generalized.

The abuser archetype is someone who will look at the product as a tool to perform harm (Fig 5.2). They may be trying to harm someone they don’t know through surveillance or anonymous harassment, or they may be trying to control, monitor, abuse, or torment someone they know personally.

The survivor archetype is someone who is being abused with the product. There are various situations to consider in terms of the archetype’s understanding of the abuse and how to put an end to it: Do they need proof of abuse they already suspect is happening, or are they unaware they’ve been targeted in the first place and need to be alerted (Fig 5.3)?

You may want to make multiple survivor archetypes to capture a range of different experiences. They may know that the abuse is happening but not be able to stop it, like when an abuser locks them out of IoT devices; or they know it’s happening but don’t know how, such as when a stalker keeps figuring out their location (Fig 5.4). Include as many of these scenarios as you need to in your survivor archetype. You’ll use these later on when you design solutions to help your survivor archetypes achieve their goals of preventing and ending abuse.

It may be useful for you to create persona-like artifacts for your archetypes, such as the three examples shown. Instead of focusing on the demographic information we often see in personas, focus on their goals. The goals of the abuser will be to carry out the specific abuse you’ve identified, while the goals of the survivor will be to prevent abuse, understand that abuse is happening, make ongoing abuse stop, or regain control over the technology that’s being used for abuse. Later, you’ll brainstorm how to prevent the abuser’s goals and assist the survivor’s goals.

And while the “abuser/survivor” model fits most cases, it doesn’t fit all, so modify it as you need to. For example, if you uncovered an issue with security, such as the ability for someone to hack into a home camera system and talk to children, the malicious hacker would get the abuser archetype and the child’s parents would get survivor archetype.

Step 3: Brainstorm problems

After creating archetypes, brainstorm novel abuse cases and safety issues. “Novel” means things not found in your research; you’re trying to identify completely new safety issues that are unique to your product or service. The goal with this step is to exhaust every effort of identifying harms your product could cause. You aren’t worrying about how to prevent the harm yet—that comes in the next step.

How could your product be used for any kind of abuse, outside of what you’ve already identified in your research? I recommend setting aside at least a few hours with your team for this process.

If you’re looking for somewhere to start, try doing a Black Mirror brainstorm. This exercise is based on the show Black Mirror, which features stories about the dark possibilities of technology. Try to figure out how your product would be used in an episode of the show—the most wild, awful, out-of-control ways it could be used for harm. When I’ve led Black Mirror brainstorms, participants usually end up having a good deal of fun (which I think is great—it’s okay to have fun when designing for safety!). I recommend time-boxing a Black Mirror brainstorm to half an hour, and then dialing it back and using the rest of the time thinking of more realistic forms of harm.

After you’ve identified as many opportunities for abuse as possible, you may still not feel confident that you’ve uncovered every potential form of harm. A healthy amount of anxiety is normal when you’re doing this kind of work. It’s common for teams designing for safety to worry, “Have we really identified every possible harm? What if we’ve missed something?” If you’ve spent at least four hours coming up with ways your product could be used for harm and have run out of ideas, go to the next step.

It’s impossible to guarantee you’ve thought of everything; instead of aiming for 100 percent assurance, recognize that you’ve taken this time and have done the best you can, and commit to continuing to prioritize safety in the future. Once your product is released, your users may identify new issues that you missed; aim to receive that feedback graciously and course-correct quickly.

Step 4: Design solutions

At this point, you should have a list of ways your product can be used for harm as well as survivor and abuser archetypes describing opposing user goals. The next step is to identify ways to design against the identified abuser’s goals and to support the survivor’s goals. This step is a good one to insert alongside existing parts of your design process where you’re proposing solutions for the various problems your research uncovered.

Some questions to ask yourself to help prevent harm and support your archetypes include:

• Can you design your product in such a way that the identified harm cannot happen in the first place? If not, what roadblocks can you put up to prevent the harm from happening?
• How can you make the victim aware that abuse is happening through your product?
• How can you help the victim understand what they need to do to make the problem stop?
• Can you identify any types of user activity that would indicate some form of harm or abuse? Could your product help the user access support?

In some products, it’s possible to proactively recognize that harm is happening. For example, a pregnancy app might be modified to allow the user to report that they were the victim of an assault, which could trigger an offer to receive resources for local and national organizations. This sort of proactiveness is not always possible, but it’s worth taking a half hour to discuss if any type of user activity would indicate some form of harm or abuse, and how your product could assist the user in receiving help in a safe manner.

That said, use caution: you don’t want to do anything that could put a user in harm’s way if their devices are being monitored. If you do offer some kind of proactive help, always make it voluntary, and think through other safety issues, such as the need to keep the user in-app in case an abuser is checking their search history. We’ll walk through a good example of this in the next chapter.

Step 5: Test for safety

The final step is to test your prototypes from the point of view of your archetypes: the person who wants to weaponize the product for harm and the victim of the harm who needs to regain control over the technology. Just like any other kind of product testing, at this point you’ll aim to rigorously test out your safety solutions so that you can identify gaps and correct them, validate that your designs will help keep your users safe, and feel more confident releasing your product into the world.

Ideally, safety testing happens along with usability testing. If you’re at a company that doesn’t do usability testing, you might be able to use safety testing to cleverly perform both; a user who goes through your design attempting to weaponize the product against someone else can also be encouraged to point out interactions or other elements of the design that don’t make sense to them.

You’ll want to conduct safety testing on either your final prototype or the actual product if it’s already been released. There’s nothing wrong with testing an existing product that wasn’t designed with safety goals in mind from the onset—“retrofitting” it for safety is a good thing to do.

Remember that testing for safety involves testing from the perspective of both an abuser and a survivor, though it may not make sense for you to do both. Alternatively, if you made multiple survivor archetypes to capture multiple scenarios, you’ll want to test from the perspective of each one.

As with other sorts of usability testing, you as the designer are most likely too close to the product and its design by this point to be a valuable tester; you know the product too well. Instead of doing it yourself, set up testing as you would with other usability testing: find someone who is not familiar with the product and its design, set the scene, give them a task, encourage them to think out loud, and observe how they attempt to complete it.

Abuser testing

The goal of this testing is to understand how easy it is for someone to weaponize your product for harm. Unlike with usability testing, you want to make it impossible, or at least difficult, for them to achieve their goal. Reference the goals in the abuser archetype you created earlier, and use your product in an attempt to achieve them.

For example, for a fitness app with GPS-enabled location features, we can imagine that the abuser archetype would have the goal of figuring out where his ex-girlfriend now lives. With this goal in mind, you’d try everything possible to figure out the location of another user who has their privacy settings enabled. You might try to see her running routes, view any available information on her profile, view anything available about her location (which she has set to private), and investigate the profiles of any other users somehow connected with her account, such as her followers.

If by the end of this you’ve managed to uncover some of her location data, despite her having set her profile to private, you know now that your product enables stalking. Your next step is to go back to step 4 and figure out how to prevent this from happening. You may need to repeat the process of designing solutions and testing them more than once.

Survivor testing

Survivor testing involves identifying how to give information and power to the survivor. It might not always make sense based on the product or context. Thwarting the attempt of an abuser archetype to stalk someone also satisfies the goal of the survivor archetype to not be stalked, so separate testing wouldn’t be needed from the survivor’s perspective.

However, there are cases where it makes sense. For example, for a smart thermostat, a survivor archetype’s goals would be to understand who or what is making the temperature change when they aren’t doing it themselves. You could test this by looking for the thermostat’s history log and checking for usernames, actions, and times; if you couldn’t find that information, you would have more work to do in step 4.

Another goal might be regaining control of the thermostat once the survivor realizes the abuser is remotely changing its settings. Your test would involve attempting to figure out how to do this: are there instructions that explain how to remove another user and change the password, and are they easy to find? This might again reveal that more work is needed to make it clear to the user how they can regain control of the device or account.

Stress testing

To make your product more inclusive and compassionate, consider adding stress testing. This concept comes from Design for Real Life by Eric Meyer and Sara Wachter-Boettcher. The authors pointed out that personas typically center people who are having a good day—but real users are often anxious, stressed out, having a bad day, or even experiencing tragedy. These are called “stress cases,” and testing your products for users in stress-case situations can help you identify places where your design lacks compassion. Design for Real Life has more details about what it looks like to incorporate stress cases into your design as well as many other great tactics for compassionate design.

Do you remember when having a great website was enough? Now, people are getting answers from Siri, Google search snippets, and mobile apps, not just our websites. Forward-thinking organizations have adopted an omnichannel content strategy, whose mission is to reach audiences across multiple digital channels and platforms.

But how do you set up a content management system (CMS) to reach your audience now and in the future? I learned the hard way that creating a content model—a definition of content types, attributes, and relationships that let people and systems understand content—with my more familiar design-system thinking would capsize my customer’s omnichannel content strategy. You can avoid that outcome by creating content models that are semantic and that also connect related content. 

I recently had the opportunity to lead the CMS implementation for a Fortune 500 company. The client was excited by the benefits of an omnichannel content strategy, including content reuse, multichannel marketing, and robot delivery—designing content to be intelligible to bots, Google knowledge panels, snippets, and voice user interfaces. 

A content model is a critical foundation for an omnichannel content strategy, and for our content to be understood by multiple systems, the model needed semantic types—types named according to their meaning instead of their presentation. Our goal was to let authors create content and reuse it wherever it was relevant. But as the project proceeded, I realized that supporting content reuse at the scale that my customer needed required the whole team to recognize a new pattern.

Despite our best intentions, we kept drawing from what we were more familiar with: design systems. Unlike web-focused content strategies, an omnichannel content strategy can’t rely on WYSIWYG tools for design and layout. Our tendency to approach the content model with our familiar design-system thinking constantly led us to veer away from one of the primary purposes of a content model: delivering content to audiences on multiple marketing channels.

Two essential principles for an effective content model

We needed to help our designers, developers, and stakeholders understand that we were doing something very different from their prior web projects, where it was natural for everyone to think about content as visual building blocks fitting into layouts. The previous approach was not only more familiar but also more intuitive—at least at first—because it made the designs feel more tangible. We discovered two principles that helped the team understand how a content model differs from the design systems that we were used to:

1. Content models must define semantics instead of layout.
2. And content models should connect content that belongs together.

Semantic content models

A semantic content model uses type and attribute names that reflect the meaning of the content, not how it will be displayed. For example, in a nonsemantic model, teams might create types like teasers, media blocks, and cards. Although these types might make it easy to lay out content, they don’t help delivery channels understand the content’s meaning, which in turn would have opened the door to the content being presented in each marketing channel. In contrast, a semantic content model uses type names like product, service, and testimonial so that each delivery channel can understand the content and use it as it sees fit. 

When you’re creating a semantic content model, a great place to start is to look over the types and properties defined by Schema.org, a community-driven resource for type definitions that are intelligible to platforms like Google search.

A semantic content model has several benefits:

• Even if your team doesn’t care about omnichannel content, a semantic content model decouples content from its presentation so that teams can evolve the website’s design without needing to refactor its content. In this way, content can withstand disruptive website redesigns. 
• A semantic content model also provides a competitive edge. By adding structured data based on Schema.org’s types and properties, a website can provide hints to help Google understand the content, display it in search snippets or knowledge panels, and use it to answer voice-interface user questions. Potential visitors could discover your content without ever setting foot in your website.
• Beyond those practical benefits, you’ll also need a semantic content model if you want to deliver omnichannel content. To use the same content in multiple marketing channels, delivery channels need to be able to understand it. For example, if your content model were to provide a list of questions and answers, it could easily be rendered on a frequently asked questions (FAQ) page, but it could also be used in a voice interface or by a bot that answers common questions.

For example, using a semantic content model for articles, events, people, and locations lets A List Apart provide cleanly structured data for search engines so that users can read the content on the website, in Google knowledge panels, and even with hypothetical voice interfaces in the future.

Content models that connect

After struggling to describe what makes a good content model, I’ve come to realize that the best models are those that are semantic and that also connect related content components (such as a FAQ item’s question and answer pair), instead of slicing up related content across disparate content components. A good content model connects content that should remain together so that multiple delivery channels can use it without needing to first put those pieces back together.

Think about writing an article or essay. An article’s meaning and usefulness depends upon its parts being kept together. Would one of the headings or paragraphs be meaningful on their own without the context of the full article? On our project, our familiar design-system thinking often led us to want to create content models that would slice content into disparate chunks to fit the web-centric layout. This had a similar impact to an article that were to have been separated from its headline. Because we were slicing content into standalone pieces based on layout, content that belonged together became difficult to manage and nearly impossible for multiple delivery channels to understand.

To illustrate, let’s look at how connecting related content applies in a real-world scenario. The design team for our customer presented a complex layout for a software product page that included multiple tabs and sections. Our instincts were to follow suit with the content model. Shouldn’t we make it as easy and as flexible as possible to add any number of tabs in the future?

Because our design-system instincts were so familiar, it felt like we had needed a content type called “tab section” so that multiple tab sections could be added to a page. Each tab section would display various types of content. One tab might provide the software’s overview or its specifications. Another tab might provide a list of resources. 

Our inclination to break down the content model into “tab section” pieces would have led to an unnecessarily complex model and a cumbersome editing experience, and it would have also created content that couldn’t have been understood by additional delivery channels. For example, how would another system have been able to tell which “tab section” referred to a product’s specifications or its resource list—would that other system have to have resorted to counting tab sections and content blocks? This would have prevented the tabs from ever being reordered, and it would have required adding logic in every other delivery channel to interpret the design system’s layout. Furthermore, if the customer were to have no longer wanted to display this content in a tab layout, it would have been tedious to migrate to a new content model to reflect the new page redesign.

We had a breakthrough when we discovered that our customer had a specific purpose in mind for each tab: it would reveal specific information such as the software product’s overview, specifications, related resources, and pricing. Once implementation began, our inclination to focus on what’s visual and familiar had obscured the intent of the designs. With a little digging, it didn’t take long to realize that the concept of tabs wasn’t relevant to the content model. The meaning of the content that they were planning to display in the tabs was what mattered.

In fact, the customer could have decided to display this content in a different way—without tabs—somewhere else. This realization prompted us to define content types for the software product based on the meaningful attributes that the customer had wanted to render on the web. There were obvious semantic attributes like name and description as well as rich attributes like screenshots, software requirements, and feature lists. The software’s product information stayed together because it wasn’t sliced across separate components like “tab sections” that were derived from the content’s presentation. Any delivery channel—including future ones—could understand and present this content.

Conclusion

In this omnichannel marketing project, we discovered that the best way to keep our content model on track was to ensure that it was semantic (with type and attribute names that reflected the meaning of the content) and that it kept content together that belonged together (instead of fragmenting it). These two concepts curtailed our temptation to shape the content model based on the design. So if you’re working on a content model to support an omnichannel content strategy—or even if you just want to make sure that Google and other interfaces understand your content—remember:

• A design system isn’t a content model. Team members may be tempted to conflate them and to make your content model mirror your design system, so you should protect the semantic value and contextual structure of the content strategy during the entire implementation process. This will let every delivery channel consume the content without needing a magic decoder ring.
• If your team is struggling to make this transition, you can still reap some of the benefits by using Schema.org–based structured data in your website. Even if additional delivery channels aren’t on the immediate horizon, the benefit to search engine optimization is a compelling reason on its own.
• Additionally, remind the team that decoupling the content model from the design will let them update the designs more easily because they won’t be held back by the cost of content migrations. They’ll be able to create new designs without the obstacle of compatibility between the design and the content, and ​they’ll be ready for the next big thing. 

By rigorously advocating for these principles, you’ll help your team treat content the way that it deserves—as the most critical asset in your user experience and the best way to connect with your audience.

Do you find yourself designing screens with only a vague idea of how the things on the screen relate to the things elsewhere in the system? Do you leave stakeholder meetings with unclear directives that often seem to contradict previous conversations? You know a better understanding of user needs would help the team get clear on what you are actually trying to accomplish, but time and budget for research is tight. When it comes to asking for more direct contact with your users, you might feel like poor Oliver Twist, timidly asking, “Please, sir, I want some more.” 

Here’s the trick. You need to get stakeholders themselves to identify high-risk assumptions and hidden complexity, so that they become just as motivated as you to get answers from users. Basically, you need to make them think it’s their idea. 

In this article, I’ll show you how to collaboratively expose misalignment and gaps in the team’s shared understanding by bringing the team together around two simple questions:

1. What are the objects?
2. What are the relationships between those objects?

A gauntlet between research and screen design

These two questions align to the first two steps of the ORCA process, which might become your new best friend when it comes to reducing guesswork. Wait, what’s ORCA?! Glad you asked.

ORCA stands for Objects, Relationships, CTAs, and Attributes, and it outlines a process for creating solid object-oriented user experiences. Object-oriented UX is my design philosophy. ORCA is an iterative methodology for synthesizing user research into an elegant structural foundation to support screen and interaction design. OOUX and ORCA have made my work as a UX designer more collaborative, effective, efficient, fun, strategic, and meaningful.

The ORCA process has four iterative rounds and a whopping fifteen steps. In each round we get more clarity on our Os, Rs, Cs, and As.

I sometimes say that ORCA is a “garbage in, garbage out” process. To ensure that the testable prototype produced in the final round actually tests well, the process needs to be fed by good research. But if you don’t have a ton of research, the beginning of the ORCA process serves another purpose: it helps you sell the need for research.

In other words, the ORCA process serves as a gauntlet between research and design. With good research, you can gracefully ride the killer whale from research into design. But without good research, the process effectively spits you back into research and with a cache of specific open questions.

Getting in the same curiosity-boat

What gets us into trouble is not what we don’t know. It’s what we know for sure that just ain’t so.

Mark Twain

The first two steps of the ORCA process—Object Discovery and Relationship Discovery—shine a spotlight on the dark, dusty corners of your team’s misalignments and any inherent complexity that’s been swept under the rug. It begins to expose what this classic comic so beautifully illustrates:

This is one reason why so many UX designers are frustrated in their job and why many projects fail. And this is also why we often can’t sell research: every decision-maker is confident in their own mental picture. 

Once we expose hidden fuzzy patches in each picture and the differences between them all, the case for user research makes itself.

But how we do this is important. However much we might want to, we can’t just tell everyone, “YOU ARE WRONG!” Instead, we need to facilitate and guide our team members to self-identify holes in their picture. When stakeholders take ownership of assumptions and gaps in understanding, BAM! Suddenly, UX research is not such a hard sell, and everyone is aboard the same curiosity-boat.

Say your users are doctors. And you have no idea how doctors use the system you are tasked with redesigning.

You might try to sell research by honestly saying: “We need to understand doctors better! What are their pain points? How do they use the current app?” But here’s the problem with that. Those questions are vague, and the answers to them don’t feel acutely actionable.

Instead, you want your stakeholders themselves to ask super-specific questions. This is more like the kind of conversation you need to facilitate. Let’s listen in:

“Wait a sec, how often do doctors share patients? Does a patient in this system have primary and secondary doctors?”

“Can a patient even have more than one primary doctor?”

“Is it a ‘primary doctor’ or just a ‘primary caregiver’… Can’t that role be a nurse practitioner?”

“No, caregivers are something else… That’s the patient’s family contacts, right?”

“So are caregivers in scope for this redesign?”

“Yeah, because if a caregiver is present at an appointment, the doctor needs to note that. Like, tag the caregiver on the note… Or on the appointment?”

Now we are getting somewhere. Do you see how powerful it can be getting stakeholders to debate these questions themselves? The diabolical goal here is to shake their confidence—gently and diplomatically.

When these kinds of questions bubble up collaboratively and come directly from the mouths of your stakeholders and decision-makers, suddenly, designing screens without knowing the answers to these questions seems incredibly risky, even silly.

If we create software without understanding the real-world information environment of our users, we will likely create software that does not align to the real-world information environment of our users. And this will, hands down, result in a more confusing, more complex, and less intuitive software product.

The two questions

But how do we get to these kinds of meaty questions diplomatically, efficiently, collaboratively, and reliably? 

We can do this by starting with those two big questions that align to the first two steps of the ORCA process:

1. What are the objects?
2. What are the relationships between those objects?

In practice, getting to these answers is easier said than done. I’m going to show you how these two simple questions can provide the outline for an Object Definition Workshop. During this workshop, these “seed” questions will blossom into dozens of specific questions and shine a spotlight on the need for more user research.

Prep work: Noun foraging

In the next section, I’ll show you how to run an Object Definition Workshop with your stakeholders (and entire cross-functional team, hopefully). But first, you need to do some prep work.

Basically, look for nouns that are particular to the business or industry of your project, and do it across at least a few sources. I call this noun foraging.

Here are just a few great noun foraging sources:

• the product’s marketing site
• the product’s competitors’ marketing sites (competitive analysis, anyone?)
• the existing product (look at labels!)
• user interview transcripts
• notes from stakeholder interviews or vision docs from stakeholders

Put your detective hat on, my dear Watson. Get resourceful and leverage what you have. If all you have is a marketing website, some screenshots of the existing legacy system, and access to customer service chat logs, then use those.

As you peruse these sources, watch for the nouns that are used over and over again, and start listing them (preferably on blue sticky notes if you’ll be creating an object map later!).

You’ll want to focus on nouns that might represent objects in your system. If you are having trouble determining if a noun might be object-worthy, remember the acronym SIP and test for:

1. Structure
2. Instances
3. Purpose

Think of a library app, for example. Is “book” an object?

Structure: can you think of a few attributes for this potential object? Title, author, publish date… Yep, it has structure. Check!

Instance: what are some examples of this potential “book” object? Can you name a few? The Alchemist, Ready Player One, Everybody Poops… OK, check!

Purpose: why is this object important to the users and business? Well, “book” is what our library client is providing to people and books are why people come to the library… Check, check, check!

As you are noun foraging, focus on capturing the nouns that have SIP. Avoid capturing components like dropdowns, checkboxes, and calendar pickers—your UX system is not your design system! Components are just the packaging for objects—they are a means to an end. No one is coming to your digital place to play with your dropdown! They are coming for the VALUABLE THINGS and what they can do with them. Those things, or objects, are what we are trying to identify.

Let’s say we work for a startup disrupting the email experience. This is how I’d start my noun foraging.

First I’d look at my own email client, which happens to be Gmail. I’d then look at Outlook and the new HEY email. I’d look at Yahoo, Hotmail…I’d even look at Slack and Basecamp and other so-called “email replacers.” I’d read some articles, reviews, and forum threads where people are complaining about email. While doing all this, I would look for and write down the nouns.

(Before moving on, feel free to go noun foraging for this hypothetical product, too, and then scroll down to see how much our lists match up. Just don’t get lost in your own emails! Come back to me!)

Drumroll, please…

Here are a few nouns I came up with during my noun foraging:

• email message
• thread
• contact
• client
• rule/automation
• email address that is not a contact?
• contact groups
• attachment
• Google doc file / other integrated file
• newsletter? (HEY treats this differently)
• saved responses and templates

Scan your list of nouns and pick out words that you are completely clueless about. In our email example, it might be client or automation. Do as much homework as you can before your session with stakeholders: google what’s googleable. But other terms might be so specific to the product or domain that you need to have a conversation about them.

Aside: here are some real nouns foraged during my own past project work that I needed my stakeholders to help me understand:

• Record Locator
• Incentive Home
• Augmented Line Item
• Curriculum-Based Measurement Probe

This is really all you need to prepare for the workshop session: a list of nouns that represent potential objects and a short list of nouns that need to be defined further.

Facilitate an Object Definition Workshop

You could actually start your workshop with noun foraging—this activity can be done collaboratively. If you have five people in the room, pick five sources, assign one to every person, and give everyone ten minutes to find the objects within their source. When the time’s up, come together and find the overlap. Affinity mapping is your friend here!

If your team is short on time and might be reluctant to do this kind of grunt work (which is usually the case) do your own noun foraging beforehand, but be prepared to show your work. I love presenting screenshots of documents and screens with all the nouns already highlighted. Bring the artifacts of your process, and start the workshop with a five-minute overview of your noun foraging journey.

HOT TIP: before jumping into the workshop, frame the conversation as a requirements-gathering session to help you better understand the scope and details of the system. You don’t need to let them know that you’re looking for gaps in the team’s understanding so that you can prove the need for more user research—that will be our little secret. Instead, go into the session optimistically, as if your knowledgeable stakeholders and PMs and biz folks already have all the answers. 

Then, let the question whack-a-mole commence.

1. What is this thing?

Want to have some real fun? At the beginning of your session, ask stakeholders to privately write definitions for the handful of obscure nouns you might be uncertain about. Then, have everyone show their cards at the same time and see if you get different definitions (you will). This is gold for exposing misalignment and starting great conversations.

As your discussion unfolds, capture any agreed-upon definitions. And when uncertainty emerges, quietly (but visibly) start an “open questions” parking lot. 😉

After definitions solidify, here’s a great follow-up:

2. Do our users know what these things are? What do users call this thing?

Stakeholder 1: They probably call email clients “apps.” But I’m not sure.

Stakeholder 2: Automations are often called “workflows,” I think. Or, maybe users think workflows are something different.

If a more user-friendly term emerges, ask the group if they can agree to use only that term moving forward. This way, the team can better align to the users’ language and mindset.

OK, moving on. 

If you have two or more objects that seem to overlap in purpose, ask one of these questions:

3. Are these the same thing? Or are these different? If they are not the same, how are they different?

You: Is a saved response the same as a template?

Stakeholder 1: Yes! Definitely.

Stakeholder 2: I don’t think so… A saved response is text with links and variables, but a template is more about the look and feel, like default fonts, colors, and placeholder images. 

Continue to build out your growing glossary of objects. And continue to capture areas of uncertainty in your “open questions” parking lot.

If you successfully determine that two similar things are, in fact, different, here’s your next follow-up question:

4. What’s the relationship between these objects?

You: Are saved responses and templates related in any way?

Stakeholder 3:  Yeah, a template can be applied to a saved response.

You, always with the follow-ups: When is the template applied to a saved response? Does that happen when the user is constructing the saved response? Or when they apply the saved response to an email? How does that actually work?

Listen. Capture uncertainty. Once the list of “open questions” grows to a critical mass, pause to start assigning questions to groups or individuals. Some questions might be for the dev team (hopefully at least one developer is in the room with you). One question might be specifically for someone who couldn’t make it to the workshop. And many questions will need to be labeled “user.” 

Do you see how we are building up to our UXR sales pitch?

5. Is this object in scope?

Your next question narrows the team’s focus toward what’s most important to your users. You can simply ask, “Are saved responses in scope for our first release?,” but I’ve got a better, more devious strategy.

By now, you should have a list of clearly defined objects. Ask participants to sort these objects from most to least important, either in small breakout groups or individually. Then, like you did with the definitions, have everyone reveal their sort order at once. Surprisingly—or not so surprisingly—it’s not unusual for the VP to rank something like “saved responses” as #2 while everyone else puts it at the bottom of the list. Try not to look too smug as you inevitably expose more misalignment.

I did this for a startup a few years ago. We posted the three groups’ wildly different sort orders on the whiteboard.

The CEO stood back, looked at it, and said, “This is why we haven’t been able to move forward in two years.”

Admittedly, it’s tragic to hear that, but as a professional, it feels pretty awesome to be the one who facilitated a watershed realization.

Once you have a good idea of in-scope, clearly defined things, this is when you move on to doing more relationship mapping.

6. Create a visual representation of the objects’ relationships

We’ve already done a bit of this while trying to determine if two things are different, but this time, ask the team about every potential relationship. For each object, ask how it relates to all the other objects. In what ways are the objects connected? To visualize all the connections, pull out your trusty boxes-and-arrows technique. Here, we are connecting our objects with verbs. I like to keep my verbs to simple “has a” and “has many” statements.

This system modeling activity brings up all sorts of new questions:

• Can a saved response have attachments?
• Can a saved response use a template? If so, if an email uses a saved response with a template, can the user override that template?
• Do users want to see all the emails they sent that included a particular attachment? For example, “show me all the emails I sent with ProfessionalImage.jpg attached. I’ve changed my professional photo and I want to alert everyone to update it.” 

Solid answers might emerge directly from the workshop participants. Great! Capture that new shared understanding. But when uncertainty surfaces, continue to add questions to your growing parking lot.

Light the fuse

You’ve positioned the explosives all along the floodgates. Now you simply have to light the fuse and BOOM. Watch the buy-in for user research flooooow.

Before your workshop wraps up, have the group reflect on the list of open questions. Make plans for getting answers internally, then focus on the questions that need to be brought before users.

Here’s your final step. Take those questions you’ve compiled for user research and discuss the level of risk associated with NOT answering them. Ask, “if we design without an answer to this question, if we make up our own answer and we are wrong, how bad might that turn out?” 

With this methodology, we are cornering our decision-makers into advocating for user research as they themselves label questions as high-risk. Sorry, not sorry. 

Now is your moment of truth. With everyone in the room, ask for a reasonable budget of time and money to conduct 6–8 user interviews focused specifically on these questions. 

HOT TIP: if you are new to UX research, please note that you’ll likely need to rephrase the questions that came up during the workshop before you present them to users. Make sure your questions are open-ended and don’t lead the user into any default answers.

Final words: Hold the screen design!

Seriously, if at all possible, do not ever design screens again without first answering these fundamental questions: what are the objects and how do they relate?

I promise you this: if you can secure a shared understanding between the business, design, and development teams before you start designing screens, you will have less heartache and save more time and money, and (it almost feels like a bonus at this point!) users will be more receptive to what you put out into the world. 

I sincerely hope this helps you win time and budget to go talk to your users and gain clarity on what you are designing before you start building screens. If you find success using noun foraging and the Object Definition Workshop, there’s more where that came from in the rest of the ORCA process, which will help prevent even more late-in-the-game scope tugs-of-war and strategy pivots. 

All the best of luck! Now go sell research!

CSS is about styling boxes. In fact, the whole web is made of boxes, from the browser viewport to elements on a page. But every once in a while a new feature comes along that makes us rethink our design approach.

Round displays, for example, make it fun to play with circular clip areas. Mobile screen notches and virtual keyboards offer challenges to best organize content that stays clear of them. And dual screen or foldable devices make us rethink how to best use available space in a number of different device postures.

These recent evolutions of the web platform made it both more challenging and more interesting to design products. They’re great opportunities for us to break out of our rectangular boxes.

I’d like to talk about a new feature similar to the above: the Window Controls Overlay for Progressive Web Apps (PWAs).

Progressive Web Apps are blurring the lines between apps and websites. They combine the best of both worlds. On one hand, they’re stable, linkable, searchable, and responsive just like websites. On the other hand, they provide additional powerful capabilities, work offline, and read files just like native apps.

As a design surface, PWAs are really interesting because they challenge us to think about what mixing web and device-native user interfaces can be. On desktop devices in particular, we have more than 40 years of history telling us what applications should look like, and it can be hard to break out of this mental model.

At the end of the day though, PWAs on desktop are constrained to the window they appear in: a rectangle with a title bar at the top.

Here’s what a typical desktop PWA app looks like:

Sure, as the author of a PWA, you get to choose the color of the title bar (using the Web Application Manifest theme_color property), but that’s about it.

What if we could think outside this box, and reclaim the real estate of the app’s entire window? Doing so would give us a chance to make our apps more beautiful and feel more integrated in the operating system.

This is exactly what the Window Controls Overlay offers. This new PWA functionality makes it possible to take advantage of the full surface area of the app, including where the title bar normally appears.

About the title bar and window controls

Let’s start with an explanation of what the title bar and window controls are.

The title bar is the area displayed at the top of an app window, which usually contains the app’s name. Window controls are the affordances, or buttons, that make it possible to minimize, maximize, or close the app’s window, and are also displayed at the top.

Window Controls Overlay removes the physical constraint of the title bar and window controls areas. It frees up the full height of the app window, enabling the title bar and window control buttons to be overlaid on top of the application’s web content. 

If you are reading this article on a desktop computer, take a quick look at other apps. Chances are they’re already doing something similar to this. In fact, the very web browser you are using to read this uses the top area to display tabs.

Spotify displays album artwork all the way to the top edge of the application window.

Microsoft Word uses the available title bar space to display the auto-save and search functionalities, and more.

The whole point of this feature is to allow you to make use of this space with your own content while providing a way to account for the window control buttons. And it enables you to offer this modified experience on a range of platforms while not adversely affecting the experience on browsers or devices that don’t support Window Controls Overlay. After all, PWAs are all about progressive enhancement, so this feature is a chance to enhance your app to use this extra space when it’s available.

Let’s use the feature

For the rest of this article, we’ll be working on a demo app to learn more about using the feature.

The demo app is called 1DIV. It’s a simple CSS playground where users can create designs using CSS and a single HTML element.

The app has two pages. The first lists the existing CSS designs you’ve created:

The second page enables you to create and edit CSS designs:

Since I’ve added a simple web manifest and service worker, we can install the app as a PWA on desktop. Here is what it looks like on macOS:

And on Windows:

Our app is looking good, but the white title bar in the first page is wasted space. In the second page, it would be really nice if the design area went all the way to the top of the app window.

Let’s use the Window Controls Overlay feature to improve this.

Enabling Window Controls Overlay

The feature is still experimental at the moment. To try it, you need to enable it in one of the supported browsers.

As of now, it has been implemented in Chromium, as a collaboration between Microsoft and Google. We can therefore use it in Chrome or Edge by going to the internal about://flags page, and enabling the Desktop PWA Window Controls Overlay flag.

Using Window Controls Overlay

To use the feature, we need to add the following display_override member to our web app’s manifest file:

{
  "name": "1DIV",
  "description": "1DIV is a mini CSS playground",
  "lang": "en-US",
  "start_url": "/",
  "theme_color": "#ffffff",
  "background_color": "#ffffff",
  "display_override": [
    "window-controls-overlay"
  ],
  "icons": [
    ...
  ]
}

On the surface, the feature is really simple to use. This manifest change is the only thing we need to make the title bar disappear and turn the window controls into an overlay.

However, to provide a great experience for all users regardless of what device or browser they use, and to make the most of the title bar area in our design, we’ll need a bit of CSS and JavaScript code.

Here is what the app looks like now:

The title bar is gone, which is what we wanted, but our logo, search field, and NEW button are partially covered by the window controls because now our layout starts at the top of the window.

It’s similar on Windows, with the difference that the close, maximize, and minimize buttons appear on the right side, grouped together with the PWA control buttons:

Using CSS to keep clear of the window controls

Along with the feature, new CSS environment variables have been introduced:

• titlebar-area-x
• titlebar-area-y
• titlebar-area-width
• titlebar-area-height

You use these variables with the CSS env() function to position your content where the title bar would have been while ensuring it won’t overlap with the window controls. In our case, we’ll use two of the variables to position our header, which contains the logo, search bar, and NEW button. 

header {
  position: absolute;
  left: env(titlebar-area-x, 0);
  width: env(titlebar-area-width, 100%);
  height: var(--toolbar-height);
}

The titlebar-area-x variable gives us the distance from the left of the viewport to where the title bar would appear, and titlebar-area-width is its width. (Remember, this is not equivalent to the width of the entire viewport, just the title bar portion, which as noted earlier, doesn’t include the window controls.)

By doing this, we make sure our content remains fully visible. We’re also defining fallback values (the second parameter in the env() function) for when the variables are not defined (such as on non-supporting browsers, or when the Windows Control Overlay feature is disabled).

Now our header adapts to its surroundings, and it doesn’t feel like the window control buttons have been added as an afterthought. The app looks a lot more like a native app.

Changing the window controls background color so it blends in

Now let’s take a closer look at our second page: the CSS playground editor.

Not great. Our CSS demo area does go all the way to the top, which is what we wanted, but the way the window controls appear as white rectangles on top of it is quite jarring.

We can fix this by changing the app’s theme color. There are a couple of ways to define it:

• PWAs can define a theme color in the web app manifest file using the theme_color manifest member. This color is then used by the OS in different ways. On desktop platforms, it is used to provide a background color to the title bar and window controls.
• Websites can use the theme-color meta tag as well. It’s used by browsers to customize the color of the UI around the web page. For PWAs, this color can override the manifest theme_color.

In our case, we can set the manifest theme_color to white to provide the right default color for our app. The OS will read this color value when the app is installed and use it to make the window controls background color white. This color works great for our main page with the list of demos.

The theme-color meta tag can be changed at runtime, using JavaScript. So we can do that to override the white with the right demo background color when one is opened.

Here is the function we’ll use:

function themeWindow(bgColor) {
  document.querySelector("meta[name=theme-color]").setAttribute('content', bgColor);
}

With this in place, we can imagine how using color and CSS transitions can produce a smooth change from the list page to the demo page, and enable the window control buttons to blend in with the rest of the app’s interface.

Dragging the window

Now, getting rid of the title bar entirely does have an important accessibility consequence: it’s much more difficult to move the application window around.

The title bar provides a sizable area for users to click and drag, but by using the Window Controls Overlay feature, this area becomes limited to where the control buttons are, and users have to very precisely aim between these buttons to move the window.

Fortunately, this can be fixed using CSS with the app-region property. This property is, for now, only supported in Chromium-based browsers and needs the -webkit- vendor prefix. 

To make any element of the app become a dragging target for the window, we can use the following: 

-webkit-app-region: drag;

It is also possible to explicitly make an element non-draggable: 

-webkit-app-region: no-drag; 

These options can be useful for us. We can make the entire header a dragging target, but make the search field and NEW button within it non-draggable so they can still be used as normal.

However, because the editor page doesn’t display the header, users wouldn’t be able to drag the window while editing code. So let’s use a different approach. We’ll create another element before our header, also absolutely positioned, and dedicated to dragging the window.

...

.drag {
  position: absolute;
  top: 0;
  width: 100%;
  height: env(titlebar-area-height, 0);
  -webkit-app-region: drag;
}

With the above code, we’re making the draggable area span the entire viewport width, and using the titlebar-area-height variable to make it as tall as what the title bar would have been. This way, our draggable area is aligned with the window control buttons as shown below.

And, now, to make sure our search field and button remain usable:

header .search,
header .new {
  -webkit-app-region: no-drag;
}

With the above code, users can click and drag where the title bar used to be. It is an area that users expect to be able to use to move windows on desktop, and we’re not breaking this expectation, which is good.

Adapting to window resize

It may be useful for an app to know both whether the window controls overlay is visible and when its size changes. In our case, if the user made the window very narrow, there wouldn’t be enough space for the search field, logo, and button to fit, so we’d want to push them down a bit.

The Window Controls Overlay feature comes with a JavaScript API we can use to do this: navigator.windowControlsOverlay.

The API provides three interesting things:

• navigator.windowControlsOverlay.visible lets us know whether the overlay is visible.
• navigator.windowControlsOverlay.getBoundingClientRect() lets us know the position and size of the title bar area.
• navigator.windowControlsOverlay.ongeometrychange lets us know when the size or visibility changes.

Let’s use this to be aware of the size of the title bar area and move the header down if it’s too narrow.

if (navigator.windowControlsOverlay) {
  navigator.windowControlsOverlay.addEventListener('geometrychange', () => {
    const { width } = navigator.windowControlsOverlay.getBoundingClientRect();
    document.body.classList.toggle('narrow', width < 250);
  });
}

In the example above, we set the narrow class on the body of the app if the title bar area is narrower than 250px. We could do something similar with a media query, but using the windowControlsOverlay API has two advantages for our use case:

• It’s only fired when the feature is supported and used; we don’t want to adapt the design otherwise.
• We get the size of the title bar area across operating systems, which is great because the size of the window controls is different on Mac and Windows. Using a media query wouldn’t make it possible for us to know exactly how much space remains.

.narrow header {
  top: env(titlebar-area-height, 0);
  left: 0;
  width: 100%;
}

Using the above CSS code, we can move our header down to stay clear of the window control buttons when the window is too narrow, and move the thumbnails down accordingly.

Thirty pixels of exciting design opportunities

Using the Window Controls Overlay feature, we were able to take our simple demo app and turn it into something that feels so much more integrated on desktop devices. Something that reaches out of the usual window constraints and provides a custom experience for its users.

In reality, this feature only gives us about 30 pixels of extra room and comes with challenges on how to deal with the window controls. And yet, this extra room and those challenges can be turned into exciting design opportunities.

More devices of all shapes and forms get invented all the time, and the web keeps on evolving to adapt to them. New features get added to the web platform to allow us, web authors, to integrate more and more deeply with those devices. From watches or foldable devices to desktop computers, we need to evolve our design approach for the web. Building for the web now lets us think outside the rectangular box.

So let’s embrace this. Let’s use the standard technologies already at our disposal, and experiment with new ideas to provide tailored experiences for all devices, all from a single codebase!

If you get a chance to try the Window Controls Overlay feature and have feedback about it, you can open issues on the spec’s repository. It’s still early in the development of this feature, and you can help make it even better. Or, you can take a look at the feature’s existing documentation, or this demo app and its source code. 

About two and a half years ago, I introduced the idea of daily ethical design. It was born out of my frustration with the many obstacles to achieving design that’s usable and equitable; protects people’s privacy, agency, and focus; benefits society; and restores nature. I argued that we need to overcome the inconveniences that prevent us from acting ethically and that we need to elevate design ethics to a more practical level by structurally integrating it into our daily work, processes, and tools.

Unfortunately, we’re still very far from this ideal. 

At the time, I didn’t know yet how to structurally integrate ethics. Yes, I had found some tools that had worked for me in previous projects, such as using checklists, assumption tracking, and “dark reality” sessions, but I didn’t manage to apply those in every project. I was still struggling for time and support, and at best I had only partially achieved a higher (moral) quality of design—which is far from my definition of structurally integrated.

I decided to dig deeper for the root causes in business that prevent us from practicing daily ethical design. Now, after much research and experimentation, I believe that I’ve found the key that will let us structurally integrate ethics. And it’s surprisingly simple! But first we need to zoom out to get a better understanding of what we’re up against.

Influence the system

Sadly, we’re trapped in a capitalistic system that reinforces consumerism and inequality, and it’s obsessed with the fantasy of endless growth. Sea levels, temperatures, and our demand for energy continue to rise unchallenged, while the gap between rich and poor continues to widen. Shareholders expect ever-higher returns on their investments, and companies feel forced to set short-term objectives that reflect this. Over the last decades, those objectives have twisted our well-intended human-centered mindset into a powerful machine that promotes ever-higher levels of consumption. When we’re working for an organization that pursues “double-digit growth” or “aggressive sales targets” (which is 99 percent of us), that’s very hard to resist while remaining human friendly. Even with our best intentions, and even though we like to say that we create solutions for people, we’re a part of the problem.

What can we do to change this?

We can start by acting on the right level of the system. Donella H. Meadows, a system thinker, once listed ways to influence a system in order of effectiveness. When you apply these to design, you get:

• At the lowest level of effectiveness, you can affect numbers such as usability scores or the number of design critiques. But none of that will change the direction of a company.
• Similarly, affecting buffers (such as team budgets), stocks (such as the number of designers), flows (such as the number of new hires), and delays (such as the time that it takes to hear about the effect of design) won’t significantly affect a company.
• Focusing instead on feedback loops such as management control, employee recognition, or design-system investments can help a company become better at achieving its objectives. But that doesn’t change the objectives themselves, which means that the organization will still work against your ethical-design ideals.
• The next level, information flows, is what most ethical-design initiatives focus on now: the exchange of ethical methods, toolkits, articles, conferences, workshops, and so on. This is also where ethical design has remained mostly theoretical. We’ve been focusing on the wrong level of the system all this time.
• Take rules, for example—they beat knowledge every time. There can be widely accepted rules, such as how finance works, or a scrum team’s definition of done. But ethical design can also be smothered by unofficial rules meant to maintain profits, often revealed through comments such as “the client didn’t ask for it” or “don’t make it too big.”
• Changing the rules without holding official power is very hard. That’s why the next level is so influential: self-organization. Experimentation, bottom-up initiatives, passion projects, self-steering teams—all of these are examples of self-organization that improve the resilience and creativity of a company. It’s exactly this diversity of viewpoints that’s needed to structurally tackle big systemic issues like consumerism, wealth inequality, and climate change.
• Yet even stronger than self-organization are objectives and metrics. Our companies want to make more money, which means that everything and everyone in the company does their best to… make the company more money. And once I realized that profit is nothing more than a measurement, I understood how crucial a very specific, defined metric can be toward pushing a company in a certain direction.

The takeaway? If we truly want to incorporate ethics into our daily design practice, we must first change the measurable objectives of the company we work for, from the bottom up.

Redefine success

Traditionally, we consider a product or service successful if it’s desirable to humans, technologically feasible, and financially viable. You tend to see these represented as equals; if you type the three words in a search engine, you’ll find diagrams of three equally sized, evenly arranged circles.

But in our hearts, we all know that the three dimensions aren’t equally weighted: it’s viability that ultimately controls whether a product will go live. So a more realistic representation might look like this:

Desirability and feasibility are the means; viability is the goal. Companies—outside of nonprofits and charities—exist to make money.

A genuinely purpose-driven company would try to reverse this dynamic: it would recognize finance for what it was intended for: a means. So both feasibility and viability are means to achieve what the company set out to achieve. It makes intuitive sense: to achieve most anything, you need resources, people, and money. (Fun fact: the Italian language knows no difference between feasibility and viability; both are simply fattibilità.)

But simply swapping viable for desirable isn’t enough to achieve an ethical outcome. Desirability is still linked to consumerism because the associated activities aim to identify what people want—whether it’s good for them or not. Desirability objectives, such as user satisfaction or conversion, don’t consider whether a product is healthy for people. They don’t prevent us from creating products that distract or manipulate people or stop us from contributing to society’s wealth inequality. They’re unsuitable for establishing a healthy balance with nature.

There’s a fourth dimension of success that’s missing: our designs also need to be ethical in the effect that they have on the world.

This is hardly a new idea. Many similar models exist, some calling the fourth dimension accountability, integrity, or responsibility. What I’ve never seen before, however, is the necessary step that comes after: to influence the system as designers and to make ethical design more practical, we must create objectives for ethical design that are achievable and inspirational. There’s no one way to do this because it highly depends on your culture, values, and industry. But I’ll give you the version that I developed with a group of colleagues at a design agency. Consider it a template to get started.

Pursue well-being, equity, and sustainability

We created objectives that address design’s effect on three levels: individual, societal, and global.

An objective on the individual level tells us what success is beyond the typical focus of usability and satisfaction—instead considering matters such as how much time and attention is required from users. We pursued well-being:

We create products and services that allow for people’s health and happiness. Our solutions are calm, transparent, nonaddictive, and nonmisleading. We respect our users’ time, attention, and privacy, and help them make healthy and respectful choices.

An objective on the societal level forces us to consider our impact beyond just the user, widening our attention to the economy, communities, and other indirect stakeholders. We called this objective equity:

We create products and services that have a positive social impact. We consider economic equality, racial justice, and the inclusivity and diversity of people as teams, users, and customer segments. We listen to local culture, communities, and those we affect.

Finally, the objective on the global level aims to ensure that we remain in balance with the only home we have as humanity. Referring to it simply as sustainability, our definition was:

We create products and services that reward sufficiency and reusability. Our solutions support the circular economy: we create value from waste, repurpose products, and prioritize sustainable choices. We deliver functionality instead of ownership, and we limit energy use.

In short, ethical design (to us) meant achieving wellbeing for each user and an equitable value distribution within society through a design that can be sustained by our living planet. When we introduced these objectives in the company, for many colleagues, design ethics and responsible design suddenly became tangible and achievable through practical—and even familiar—actions.

Measure impact 

But defining these objectives still isn’t enough. What truly caught the attention of senior management was the fact that we created a way to measure every design project’s well-being, equity, and sustainability.

This overview lists example metrics that you can use as you pursue well-being, equity, and sustainability:

There’s a lot of power in measurement. As the saying goes, what gets measured gets done. Donella Meadows once shared this example:

“If the desired system state is national security, and that is defined as the amount of money spent on the military, the system will produce military spending. It may or may not produce national security.”

This phenomenon explains why desirability is a poor indicator of success: it’s typically defined as the increase in customer satisfaction, session length, frequency of use, conversion rate, churn rate, download rate, and so on. But none of these metrics increase the health of people, communities, or ecosystems. What if instead we measured success through metrics for (digital) well-being, such as (reduced) screen time or software energy consumption?

There’s another important message here. Even if we set an objective to build a calm interface, if we were to choose the wrong metric for calmness—say, the number of interface elements—we could still end up with a screen that induces anxiety. Choosing the wrong metric can completely undo good intentions. 

Additionally, choosing the right metric is enormously helpful in focusing the design team. Once you go through the exercise of choosing metrics for our objectives, you’re forced to consider what success looks like concretely and how you can prove that you’ve reached your ethical objectives. It also forces you to consider what we as designers have control over: what can I include in my design or change in my process that will lead to the right type of success? The answer to this question brings a lot of clarity and focus.

And finally, it’s good to remember that traditional businesses run on measurements, and managers love to spend much time discussing charts (ideally hockey-stick shaped)—especially if they concern profit, the one-above-all of metrics. For good or ill, to improve the system, to have a serious discussion about ethical design with managers, we’ll need to speak that business language.

Practice daily ethical design

Once you’ve defined your objectives and you have a reasonable idea of the potential metrics for your design project, only then do you have a chance to structurally practice ethical design. It “simply” becomes a matter of using your creativity and choosing from all the knowledge and toolkits already available to you.

I think this is quite exciting! It opens a whole new set of challenges and considerations for the design process. Should you go with that energy-consuming video or would a simple illustration be enough? Which typeface is the most calm and inclusive? Which new tools and methods do you use? When is the website’s end of life? How can you provide the same service while requiring less attention from users? How do you make sure that those who are affected by decisions are there when those decisions are made? How can you measure our effects?

The redefinition of success will completely change what it means to do good design.

There is, however, a final piece of the puzzle that’s missing: convincing your client, product owner, or manager to be mindful of well-being, equity, and sustainability. For this, it’s essential to engage stakeholders in a dedicated kickoff session.

Kick it off or fall back to status quo

The kickoff is the most important meeting that can be so easy to forget to include. It consists of two major phases: 1) the alignment of expectations, and 2) the definition of success.

In the first phase, the entire (design) team goes over the project brief and meets with all the relevant stakeholders. Everyone gets to know one another and express their expectations on the outcome and their contributions to achieving it. Assumptions are raised and discussed. The aim is to get on the same level of understanding and to in turn avoid preventable miscommunications and surprises later in the project.

For example, for a recent freelance project that aimed to design a digital platform that facilitates US student advisors’ documentation and communication, we conducted an online kickoff with the client, a subject-matter expert, and two other designers. We used a combination of canvases on Miro: one with questions from “Manual of Me” (to get to know each other), a Team Canvas (to express expectations), and a version of the Project Canvas to align on scope, timeline, and other practical matters.

The above is the traditional purpose of a kickoff. But just as important as expressing expectations is agreeing on what success means for the project—in terms of desirability, viability, feasibility, and ethics. What are the objectives in each dimension?

Agreement on what success means at such an early stage is crucial because you can rely on it for the remainder of the project. If, for example, the design team wants to build an inclusive app for a diverse user group, they can raise diversity as a specific success criterion during the kickoff. If the client agrees, the team can refer back to that promise throughout the project. “As we agreed in our first meeting, having a diverse user group that includes A and B is necessary to build a successful product. So we do activity X and follow research process Y.” Compare those odds to a situation in which the team didn’t agree to that beforehand and had to ask for permission halfway through the project. The client might argue that that came on top of the agreed scope—and she’d be right.

In the case of this freelance project, to define success I prepared a round canvas that I call the Wheel of Success. It consists of an inner ring, meant to capture ideas for objectives, and a set of outer rings, meant to capture ideas on how to measure those objectives. The rings are divided into five dimensions of successful design: healthy, equitable, sustainable, desirable, feasible, and viable.

We went through each dimension, writing down ideas on digital sticky notes. Then we discussed our ideas and verbally agreed on the most important ones. For example, our client agreed that sustainability and progressive enhancement are important success criteria for the platform. And the subject-matter expert emphasized the importance of including students from low-income and disadvantaged groups in the design process.

After the kickoff, we summarized our ideas and shared understanding in a project brief that captured these aspects:

• the project’s origin and purpose: why are we doing this project?
• the problem definition: what do we want to solve?
• the concrete goals and metrics for each success dimension: what do we want to achieve?
• the scope, process, and role descriptions: how will we achieve it?

With such a brief in place, you can use the agreed-upon objectives and concrete metrics as a checklist of success, and your design team will be ready to pursue the right objective—using the tools, methods, and metrics at their disposal to achieve ethical outcomes.

Conclusion

Over the past year, quite a few colleagues have asked me, “Where do I start with ethical design?” My answer has always been the same: organize a session with your stakeholders to (re)define success. Even though you might not always be 100 percent successful in agreeing on goals that cover all responsibility objectives, that beats the alternative (the status quo) every time. If you want to be an ethical, responsible designer, there’s no skipping this step.

To be even more specific: if you consider yourself a strategic designer, your challenge is to define ethical objectives, set the right metrics, and conduct those kick-off sessions. If you consider yourself a system designer, your starting point is to understand how your industry contributes to consumerism and inequality, understand how finance drives business, and brainstorm which levers are available to influence the system on the highest level. Then redefine success to create the space to exercise those levers.

And for those who consider themselves service designers or UX designers or UI designers: if you truly want to have a positive, meaningful impact, stay away from the toolkits and meetups and conferences for a while. Instead, gather your colleagues and define goals for well-being, equity, and sustainability through design. Engage your stakeholders in a workshop and challenge them to think of ways to achieve and measure those ethical goals. Take their input, make it concrete and visible, ask for their agreement, and hold them to it.

Otherwise, I’m genuinely sorry to say, you’re wasting your precious time and creative energy.

Of course, engaging your stakeholders in this way can be uncomfortable. Many of my colleagues expressed doubts such as “What will the client think of this?,” “Will they take me seriously?,” and “Can’t we just do it within the design team instead?” In fact, a product manager once asked me why ethics couldn’t just be a structured part of the design process—to just do it without spending the effort to define ethical objectives. It’s a tempting idea, right? We wouldn’t have to have difficult discussions with stakeholders about what values or which key-performance indicators to pursue. It would let us focus on what we like and do best: designing.

But as systems theory tells us, that’s not enough. For those of us who aren’t from marginalized groups and have the privilege to be able to speak up and be heard, that uncomfortable space is exactly where we need to be if we truly want to make a difference. We can’t remain within the design-for-designers bubble, enjoying our privileged working-from-home situation, disconnected from the real world out there. For those of us who have the possibility to speak up and be heard: if we solely keep talking about ethical design and it remains at the level of articles and toolkits—we’re not designing ethically. It’s just theory. We need to actively engage our colleagues and clients by challenging them to redefine success in business.

With a bit of courage, determination, and focus, we can break out of this cage that finance and business-as-usual have built around us and become facilitators of a new type of business that can see beyond financial value. We just need to agree on the right objectives at the start of each design project, find the right metrics, and realize that we already have everything that we need to get started. That’s what it means to do daily ethical design.

For their inspiration and support over the years, I would like to thank Emanuela Cozzi Schettini, José Gallegos, Annegret Bönemann, Ian Dorr, Vera Rademaker, Virginia Rispoli, Cecilia Scolaro, Rouzbeh Amini, and many others.

The mobile-first design methodology is great—it focuses on what really matters to the user, it’s well-practiced, and it’s been a common design pattern for years. So developing your CSS mobile-first should also be great, too…right? 

Well, not necessarily. Classic mobile-first CSS development is based on the principle of overwriting style declarations: you begin your CSS with default style declarations, and overwrite and/or add new styles as you add breakpoints with min-width media queries for larger viewports (for a good overview see “What is Mobile First CSS and Why Does It Rock?”). But all those exceptions create complexity and inefficiency, which in turn can lead to an increased testing effort and a code base that’s harder to maintain. Admit it—how many of us willingly want that?

On your own projects, mobile-first CSS may yet be the best tool for the job, but first you need to evaluate just how appropriate it is in light of the visual design and user interactions you’re working on. To help you get started, here’s how I go about tackling the factors you need to watch for, and I’ll discuss some alternate solutions if mobile-first doesn’t seem to suit your project.

Advantages of mobile-first

Some of the things to like with mobile-first CSS development—and why it’s been the de facto development methodology for so long—make a lot of sense:

Development hierarchy. One thing you undoubtedly get from mobile-first is a nice development hierarchy—you just focus on the mobile view and get developing. 

Tried and tested. It’s a tried and tested methodology that’s worked for years for a reason: it solves a problem really well.

Prioritizes the mobile view. The mobile view is the simplest and arguably the most important, as it encompasses all the key user journeys, and often accounts for a higher proportion of user visits (depending on the project). 

Prevents desktop-centric development. As development is done using desktop computers, it can be tempting to initially focus on the desktop view. But thinking about mobile from the start prevents us from getting stuck later on; no one wants to spend their time retrofitting a desktop-centric site to work on mobile devices!

Disadvantages of mobile-first

Setting style declarations and then overwriting them at higher breakpoints can lead to undesirable ramifications:

More complexity. The farther up the breakpoint hierarchy you go, the more unnecessary code you inherit from lower breakpoints. 

Higher CSS specificity. Styles that have been reverted to their browser default value in a class name declaration now have a higher specificity. This can be a headache on large projects when you want to keep the CSS selectors as simple as possible.

Requires more regression testing. Changes to the CSS at a lower view (like adding a new style) requires all higher breakpoints to be regression tested.

The browser can’t prioritize CSS downloads. At wider breakpoints, classic mobile-first min-width media queries don’t leverage the browser’s capability to download CSS files in priority order.

The problem of property value overrides

There is nothing inherently wrong with overwriting values; CSS was designed to do just that. Still, inheriting incorrect values is unhelpful and can be burdensome and inefficient. It can also lead to increased style specificity when you have to overwrite styles to reset them back to their defaults, something that may cause issues later on, especially if you are using a combination of bespoke CSS and utility classes. We won’t be able to use a utility class for a style that has been reset with a higher specificity.

With this in mind, I’m developing CSS with a focus on the default values much more these days. Since there’s no specific order, and no chains of specific values to keep track of, this frees me to develop breakpoints simultaneously. I concentrate on finding common styles and isolating the specific exceptions in closed media query ranges (that is, any range with a max-width set). 

This approach opens up some opportunities, as you can look at each breakpoint as a clean slate. If a component’s layout looks like it should be based on Flexbox at all breakpoints, it’s fine and can be coded in the default style sheet. But if it looks like Grid would be much better for large screens and Flexbox for mobile, these can both be done entirely independently when the CSS is put into closed media query ranges. Also, developing simultaneously requires you to have a good understanding of any given component in all breakpoints up front. This can help surface issues in the design earlier in the development process. We don’t want to get stuck down a rabbit hole building a complex component for mobile, and then get the designs for desktop and find they are equally complex and incompatible with the HTML we created for the mobile view! 

Though this approach isn’t going to suit everyone, I encourage you to give it a try. There are plenty of tools out there to help with concurrent development, such as Responsively App, Blisk, and many others. 

Having said that, I don’t feel the order itself is particularly relevant. If you are comfortable with focusing on the mobile view, have a good understanding of the requirements for other breakpoints, and prefer to work on one device at a time, then by all means stick with the classic development order. The important thing is to identify common styles and exceptions so you can put them in the relevant stylesheet—a sort of manual tree-shaking process! Personally, I find this a little easier when working on a component across breakpoints, but that’s by no means a requirement.

Closed media query ranges in practice 

In classic mobile-first CSS we overwrite the styles, but we can avoid this by using media query ranges. To illustrate the difference (I’m using SCSS for brevity), let’s assume there are three visual designs: 

• smaller than 768
• from 768 to below 1024
• 1024 and anything larger 

Take a simple example where a block-level element has a default padding of “20px,” which is overwritten at tablet to be “40px” and set back to “20px” on desktop.

.my-block {
  padding: 20px;
  @media (min-width: 768px) {
    padding: 40px;
  }
  @media (min-width: 1024px) {
    padding: 20px;
  }
}

.my-block {
  padding: 20px;
  @media (min-width: 768px) and (max-width: 1023.98px) {
    padding: 40px;
  }
}

The subtle difference is that the mobile-first example sets the default padding to “20px” and then overwrites it at each breakpoint, setting it three times in total. In contrast, the second example sets the default padding to “20px” and only overrides it at the relevant breakpoint where it isn’t the default value (in this instance, tablet is the exception).

The goal is to: 

• Only set styles when needed. 
• Not set them with the expectation of overwriting them later on, again and again. 

To this end, closed media query ranges are our best friend. If we need to make a change to any given view, we make it in the CSS media query range that applies to the specific breakpoint. We’ll be much less likely to introduce unwanted alterations, and our regression testing only needs to focus on the breakpoint we have actually edited. 

Taking the above example, if we find that .my-block spacing on desktop is already accounted for by the margin at that breakpoint, and since we want to remove the padding altogether, we could do this by setting the mobile padding in a closed media query range.

.my-block {
  @media (max-width: 767.98px) {
    padding: 20px;
  }
  @media (min-width: 768px) and (max-width: 1023.98px) {
    padding: 40px;
  }
}

The browser default padding for our block is “0,” so instead of adding a desktop media query and using unset or “0” for the padding value (which we would need with mobile-first), we can wrap the mobile padding in a closed media query (since it is now also an exception) so it won’t get picked up at wider breakpoints. At the desktop breakpoint, we won’t need to set any padding style, as we want the browser default value.

Bundling versus separating the CSS

Back in the day, keeping the number of requests to a minimum was very important due to the browser’s limit of concurrent requests (typically around six). As a consequence, the use of image sprites and CSS bundling was the norm, with all the CSS being downloaded in one go, as one stylesheet with highest priority. 

With HTTP/2 and HTTP/3 now on the scene, the number of requests is no longer the big deal it used to be. This allows us to separate the CSS into multiple files by media query. The clear benefit of this is the browser can now request the CSS it currently needs with a higher priority than the CSS it doesn’t. This is more performant and can reduce the overall time page rendering is blocked.

Which HTTP version are you using?

To determine which version of HTTP you’re using, go to your website and open your browser’s dev tools. Next, select the Network tab and make sure the Protocol column is visible. If “h2” is listed under Protocol, it means HTTP/2 is being used. 

Note: to view the Protocol in your browser’s dev tools, go to the Network tab, reload your page, right-click any column header (e.g., Name), and check the Protocol column.

Also, if your site is still using HTTP/1...WHY?!! What are you waiting for? There is excellent user support for HTTP/2.

Splitting the CSS

Separating the CSS into individual files is a worthwhile task. Linking the separate CSS files using the relevant media attribute allows the browser to identify which files are needed immediately (because they’re render-blocking) and which can be deferred. Based on this, it allocates each file an appropriate priority.

In the following example of a website visited on a mobile breakpoint, we can see the mobile and default CSS are loaded with “Highest” priority, as they are currently needed to render the page. The remaining CSS files (print, tablet, and desktop) are still downloaded in case they’ll be needed later, but with “Lowest” priority. 

With bundled CSS, the browser will have to download the CSS file and parse it before rendering can start.

While, as noted, with the CSS separated into different files linked and marked up with the relevant media attribute, the browser can prioritize the files it currently needs. Using closed media query ranges allows the browser to do this at all widths, as opposed to classic mobile-first min-width queries, where the desktop browser would have to download all the CSS with Highest priority. We can’t assume that desktop users always have a fast connection. For instance, in many rural areas, internet connection speeds are still slow. 

The media queries and number of separate CSS files will vary from project to project based on project requirements, but might look similar to the example below.

Depending on the project’s deployment strategy, a change to one file (mobile.css, for example) would only require the QA team to regression test on devices in that specific media query range. Compare that to the prospect of deploying the single bundled site.css file, an approach that would normally trigger a full regression test.

Moving on

The uptake of mobile-first CSS was a really important milestone in web development; it has helped front-end developers focus on mobile web applications, rather than developing sites on desktop and then attempting to retrofit them to work on other devices.

I don’t think anyone wants to return to that development model again, but it’s important we don’t lose sight of the issue it highlighted: that things can easily get convoluted and less efficient if we prioritize one particular device—any device—over others. For this reason, focusing on the CSS in its own right, always mindful of what is the default setting and what’s an exception, seems like the natural next step. I’ve started noticing small simplifications in my own CSS, as well as other developers’, and that testing and maintenance work is also a bit more simplified and productive. 

In general, simplifying CSS rule creation whenever we can is ultimately a cleaner approach than going around in circles of overrides. But whichever methodology you choose, it needs to suit the project. Mobile-first may—or may not—turn out to be the best choice for what’s involved, but first you need to solidly understand the trade-offs you’re stepping into.

As a UX professional in today’s data-driven landscape, it’s increasingly likely that you’ve been asked to design a personalized digital experience, whether it’s a public website, user portal, or native application. Yet while there continues to be no shortage of marketing hype around personalization platforms, we still have very few standardized approaches for implementing personalized UX.

That’s where we come in. After completing dozens of personalization projects over the past few years, we gave ourselves a goal: could you create a holistic personalization framework specifically for UX practitioners? The Personalization Pyramid is a designer-centric model for standing up human-centered personalization programs, spanning data, segmentation, content delivery, and overall goals. By using this approach, you will be able to understand the core components of a contemporary, UX-driven personalization program (or at the very least know enough to get started). 

Getting Started

For the sake of this article, we’ll assume you’re already familiar with the basics of digital personalization. A good overview can be found here: Website Personalization Planning. While UX projects in this area can take on many different forms, they often stem from similar starting points.      

Common scenarios for starting a personalization project:

• Your organization or client purchased a content management system (CMS) or marketing automation platform (MAP) or related technology that supports personalization
• The CMO, CDO, or CIO has identified personalization as a goal
• Customer data is disjointed or ambiguous
• You are running some isolated targeting campaigns or A/B testing
• Stakeholders disagree on personalization approach
• Mandate of customer privacy rules (e.g. GDPR) requires revisiting existing user targeting practices

Regardless of where you begin, a successful personalization program will require the same core building blocks. We’ve captured these as the “levels” on the pyramid. Whether you are a UX designer, researcher, or strategist, understanding the core components can help make your contribution successful.  

From top to bottom, the levels include:

1. North Star: What larger strategic objective is driving the personalization program? 
2. Goals: What are the specific, measurable outcomes of the program? 
3. Touchpoints: Where will the personalized experience be served?
4. Contexts and Campaigns: What personalization content will the user see?
5. User Segments: What constitutes a unique, usable audience? 
6. Actionable Data: What reliable and authoritative data is captured by our technical platform to drive personalization?  
7. Raw Data: What wider set of data is conceivably available (already in our setting) allowing you to personalize?

We’ll go through each of these levels in turn. To help make this actionable, we created an accompanying deck of cards to illustrate specific examples from each level. We’ve found them helpful in personalization brainstorming sessions, and will include examples for you here.

Starting at the Top

The components of the pyramid are as follows:

North Star

A north star is what you are aiming for overall with your personalization program (big or small). The North Star defines the (one) overall mission of the personalization program. What do you wish to accomplish? North Stars cast a shadow. The bigger the star, the bigger the shadow. Example of North Starts might include: 

1. Function: Personalize based on basic user inputs. Examples: “Raw” notifications, basic search results, system user settings and configuration options, general customization, basic optimizations
2. Feature: Self-contained personalization componentry. Examples: “Cooked” notifications, advanced optimizations (geolocation), basic dynamic messaging, customized modules, automations, recommenders
3. Experience: Personalized user experiences across multiple interactions and user flows. Examples: Email campaigns, landing pages, advanced messaging (i.e. C2C chat) or conversational interfaces, larger user flows and content-intensive optimizations (localization).
4. Product: Highly differentiating personalized product experiences. Examples: Standalone, branded experiences with personalization at their core, like the “algotorial” playlists by Spotify such as Discover Weekly.

Goals

As in any good UX design, personalization can help accelerate designing with customer intentions. Goals are the tactical and measurable metrics that will prove the overall program is successful. A good place to start is with your current analytics and measurement program and metrics you can benchmark against. In some cases, new goals may be appropriate. The key thing to remember is that personalization itself is not a goal, rather it is a means to an end. Common goals include:

• Conversion
• Time on task
• Net promoter score (NPS)
• Customer satisfaction 

Touchpoints

Touchpoints are where the personalization happens. As a UX designer, this will be one of your largest areas of responsibility. The touchpoints available to you will depend on how your personalization and associated technology capabilities are instrumented, and should be rooted in improving a user’s experience at a particular point in the journey. Touchpoints can be multi-device (mobile, in-store, website) but also more granular (web banner, web pop-up etc.). Here are some examples:

Channel-level Touchpoints

• Email: Role
• Email: Time of open
• In-store display (JSON endpoint)
• Native app
• Search

Wireframe-level Touchpoints

• Web overlay
• Web alert bar
• Web banner
• Web content block
• Web menu

If you’re designing for web interfaces, for example, you will likely need to include personalized “zones” in your wireframes. The content for these can be presented programmatically in touchpoints based on our next step, contexts and campaigns.

Contexts and Campaigns

Once you’ve outlined some touchpoints, you can consider the actual personalized content a user will receive. Many personalization tools will refer to these as “campaigns” (so, for example, a campaign on a web banner for new visitors to the website). These will programmatically be shown at certain touchpoints to certain user segments, as defined by user data. At this stage, we find it helpful to consider two separate models: a context model and a content model. The context helps you consider the level of engagement of the user at the personalization moment, for example a user casually browsing information vs. doing a deep-dive. Think of it in terms of information retrieval behaviors. The content model can then help you determine what type of personalization to serve based on the context (for example, an “Enrich” campaign that shows related articles may be a suitable supplement to extant content).

Personalization Context Model:

1. Browse
2. Skim
3. Nudge
4. Feast

Personalization Content Model:

1. Alert
2. Make Easier
3. Cross-Sell
4. Enrich

We’ve written extensively about each of these models elsewhere, so if you’d like to read more you can check out Colin’s Personalization Content Model and Jeff’s Personalization Context Model. 

User Segments

User segments can be created prescriptively or adaptively, based on user research (e.g. via rules and logic tied to set user behaviors or via A/B testing). At a minimum you will likely need to consider how to treat the unknown or first-time visitor, the guest or returning visitor for whom you may have a stateful cookie (or equivalent post-cookie identifier), or the authenticated visitor who is logged in. Here are some examples from the personalization pyramid:

• Unknown
• Guest
• Authenticated
• Default
• Referred
• Role
• Cohort
• Unique ID

Actionable Data

Every organization with any digital presence has data. It’s a matter of asking what data you can ethically collect on users, its inherent reliability and value, as to how can you use it (sometimes known as “data activation.”) Fortunately, the tide is turning to first-party data: a recent study by Twilio estimates some 80% of businesses are using at least some type of first-party data to personalize the customer experience. 

First-party data represents multiple advantages on the UX front, including being relatively simple to collect, more likely to be accurate, and less susceptible to the “creep factor” of third-party data. So a key part of your UX strategy should be to determine what the best form of data collection is on your audiences. Here are some examples:

There is a progression of profiling when it comes to recognizing and making decisioning about different audiences and their signals. It tends to move towards more granular constructs about smaller and smaller cohorts of users as time and confidence and data volume grow.

While some combination of implicit / explicit data is generally a prerequisite for any implementation (more commonly referred to as first party and third-party data) ML efforts are typically not cost-effective directly out of the box. This is because a strong data backbone and content repository is a prerequisite for optimization. But these approaches should be considered as part of the larger roadmap and may indeed help accelerate the organization’s overall progress. Typically at this point you will partner with key stakeholders and product owners to design a profiling model. The profiling model includes defining approach to configuring profiles, profile keys, profile cards and pattern cards. A multi-faceted approach to profiling which makes it scalable.

Pulling it Together

While the cards comprise the starting point to an inventory of sorts (we provide blanks for you to tailor your own), a set of potential levers and motivations for the style of personalization activities you aspire to deliver, they are more valuable when thought of in a grouping. 

In assembling a card “hand”, one can begin to trace the entire trajectory from leadership focus down through a strategic and tactical execution. It is also at the heart of the way both co-authors have conducted workshops in assembling a program backlog—which is a fine subject for another article.

In the meantime, what is important to note is that each colored class of card is helpful to survey in understanding the range of choices potentially at your disposal, it is threading through and making concrete decisions about for whom this decisioning will be made: where, when, and how.

Lay Down Your Cards

Any sustainable personalization strategy must consider near, mid and long-term goals. Even with the leading CMS platforms like Sitecore and Adobe or the most exciting composable CMS DXP out there, there is simply no “easy button” wherein a personalization program can be stood up and immediately view meaningful results. That said, there is a common grammar to all personalization activities, just like every sentence has nouns and verbs. These cards attempt to map that territory.

Humility, a designer’s essential value—that has a nice ring to it. What about humility, an office manager’s essential value? Or a dentist’s? Or a librarian’s? They all sound great. When humility is our guiding light, the path is always open for fulfillment, evolution, connection, and engagement. In this chapter, we’re going to talk about why.

That said, this is a book for designers, and to that end, I’d like to start with a story—well, a journey, really. It’s a personal one, and I’m going to make myself a bit vulnerable along the way. I call it:

The Tale of Justin’s Preposterous Pate

When I was coming out of art school, a long-haired, goateed neophyte, print was a known quantity to me; design on the web, however, was rife with complexities to navigate and discover, a problem to be solved. Though I had been formally trained in graphic design, typography, and layout, what fascinated me was how these traditional skills might be applied to a fledgling digital landscape. This theme would ultimately shape the rest of my career.

So rather than graduate and go into print like many of my friends, I devoured HTML and JavaScript books into the wee hours of the morning and taught myself how to code during my senior year. I wanted—nay, needed—to better understand the underlying implications of what my design decisions would mean once rendered in a browser.

The late ’90s and early 2000s were the so-called “Wild West” of web design. Designers at the time were all figuring out how to apply design and visual communication to the digital landscape. What were the rules? How could we break them and still engage, entertain, and convey information? At a more macro level, how could my values, inclusive of humility, respect, and connection, align in tandem with that? I was hungry to find out.

Though I’m talking about a different era, those are timeless considerations between non-career interactions and the world of design. What are your core passions, or values, that transcend medium? It’s essentially the same concept we discussed earlier on the direct parallels between what fulfills you, agnostic of the tangible or digital realms; the core themes are all the same.

First within tables, animated GIFs, Flash, then with Web Standards, divs, and CSS, there was personality, raw unbridled creativity, and unique means of presentment that often defied any semblance of a visible grid. Splash screens and “browser requirement” pages aplenty. Usability and accessibility were typically victims of such a creation, but such paramount facets of any digital design were largely (and, in hindsight, unfairly) disregarded at the expense of experimentation.

For example, this iteration of my personal portfolio site (“the pseudoroom”) from that era was experimental, if not a bit heavy- handed, in the visual communication of the concept of a living sketchbook. Very skeuomorphic. I collaborated with fellow designer and dear friend Marc Clancy (now a co-founder of the creative project organizing app Milanote) on this one, where we’d first sketch and then pass a Photoshop file back and forth to trick things out and play with varied user interactions. Then, I’d break it down and code it into a digital layout.

Along with design folio pieces, the site also offered free downloads for Mac OS customizations: desktop wallpapers that were effectively design experimentation, custom-designed typefaces, and desktop icons.

From around the same time, GUI Galaxy was a design, pixel art, and Mac-centric news portal some graphic designer friends and I conceived, designed, developed, and deployed.

Design news portals were incredibly popular during this period, featuring (what would now be considered) Tweet-size, small-format snippets of pertinent news from the categories I previously mentioned. If you took Twitter, curated it to a few categories, and wrapped it in a custom-branded experience, you’d have a design news portal from the late 90s / early 2000s.

We as designers had evolved and created a bandwidth-sensitive, web standards award-winning, much more accessibility-conscious website. Still ripe with experimentation, yet more mindful of equitable engagement. You can see a couple of content panes here, noting general news (tech, design) and Mac-centric news below. We also offered many of the custom downloads I cited before as present on my folio site but branded and themed to GUI Galaxy.

The site’s backbone was a homegrown CMS, with the presentation layer consisting of global design + illustration + news author collaboration. And the collaboration effort here, in addition to experimentation on a ‘brand’ and content delivery, was hitting my core. We were designing something bigger than any single one of us and connecting with a global audience.

Collaboration and connection transcend medium in their impact, immensely fulfilling me as a designer.

Now, why am I taking you down this trip of design memory lane? Two reasons.

First, there’s a reason for the nostalgia for that design era (the “Wild West” era, as I called it earlier): the inherent exploration, personality, and creativity that saturated many design portals and personal portfolio sites. Ultra-finely detailed pixel art UI, custom illustration, bespoke vector graphics, all underpinned by a strong design community.

Today’s web design has been in a period of stagnation. I suspect there’s a strong chance you’ve seen a site whose structure looks something like this: a hero image / banner with text overlaid, perhaps with a lovely rotating carousel of images (laying the snark on heavy there), a call to action, and three columns of sub-content directly beneath. Maybe an icon library is employed with selections that vaguely relate to their respective content.

Design, as it’s applied to the digital landscape, is in dire need of thoughtful layout, typography, and visual engagement that goes hand-in-hand with all the modern considerations we now know are paramount: usability. Accessibility. Load times and bandwidth- sensitive content delivery. A responsive presentation that meets human beings wherever they’re engaging from. We must be mindful of, and respectful toward, those concerns—but not at the expense of creativity of visual communication or via replicating cookie-cutter layouts.

Pixel Problems

Websites during this period were often designed and built on Macs whose OS and desktops looked something like this. This is Mac OS 7.5, but 8 and 9 weren’t that different.

Desktop icons fascinated me: how could any single one, at any given point, stand out to get my attention? In this example, the user’s desktop is tidy, but think of a more realistic example with icon pandemonium. Or, say an icon was part of a larger system grouping (fonts, extensions, control panels)—how did it also maintain cohesion amongst a group?

These were 32 x 32 pixel creations, utilizing a 256-color palette, designed pixel-by-pixel as mini mosaics. To me, this was the embodiment of digital visual communication under such ridiculous constraints. And often, ridiculous restrictions can yield the purification of concept and theme.

So I began to research and do my homework. I was a student of this new medium, hungry to dissect, process, discover, and make it my own.

Expanding upon the notion of exploration, I wanted to see how I could push the limits of a 32×32 pixel grid with that 256-color palette. Those ridiculous constraints forced a clarity of concept and presentation that I found incredibly appealing. The digital gauntlet had been tossed, and that challenge fueled me. And so, in my dorm room into the wee hours of the morning, I toiled away, bringing conceptual sketches into mini mosaic fruition.

These are some of my creations, utilizing the only tool available at the time to create icons called ResEdit. ResEdit was a clunky, built-in Mac OS utility not really made for exactly what we were using it for. At the core of all of this work: Research. Challenge. Problem- solving. Again, these core connection-based values are agnostic of medium.

There’s one more design portal I want to talk about, which also serves as the second reason for my story to bring this all together.

This is K10k, short for Kaliber 1000. K10k was founded in 1998 by Michael Schmidt and Toke Nygaard, and was the design news portal on the web during this period. With its pixel art-fueled presentation, ultra-focused care given to every facet and detail, and with many of the more influential designers of the time who were invited to be news authors on the site, well… it was the place to be, my friend. With respect where respect is due, GUI Galaxy’s concept was inspired by what these folks were doing.

For my part, the combination of my web design work and pixel art exploration began to get me some notoriety in the design scene. Eventually, K10k noticed and added me as one of their very select group of news authors to contribute content to the site.

Amongst my personal work and side projects—and now with this inclusion—in the design community, this put me on the map. My design work also began to be published in various printed collections, in magazines domestically and overseas, and featured on other design news portals. With that degree of success while in my early twenties, something else happened:

I evolved—devolved, really—into a colossal asshole (and in just about a year out of art school, no less). The press and the praise became what fulfilled me, and they went straight to my head. They inflated my ego. I actually felt somewhat superior to my fellow designers.

The casualties? My design stagnated. Its evolution—my evolution— stagnated.

I felt so supremely confident in my abilities that I effectively stopped researching and discovering. When previously sketching concepts or iterating ideas in lead was my automatic step one, I instead leaped right into Photoshop. I drew my inspiration from the smallest of sources (and with blinders on). Any critique of my work from my peers was often vehemently dismissed. The most tragic loss: I had lost touch with my values.

My ego almost cost me some of my friendships and burgeoning professional relationships. I was toxic in talking about design and in collaboration. But thankfully, those same friends gave me a priceless gift: candor. They called me out on my unhealthy behavior.

Admittedly, it was a gift I initially did not accept but ultimately was able to deeply reflect upon. I was soon able to accept, and process, and course correct. The realization laid me low, but the re-awakening was essential. I let go of the “reward” of adulation and re-centered upon what stoked the fire for me in art school. Most importantly: I got back to my core values.

Always Students

Following that short-term regression, I was able to push forward in my personal design and career. And I could self-reflect as I got older to facilitate further growth and course correction as needed.

As an example, let’s talk about the Large Hadron Collider. The LHC was designed “to help answer some of the fundamental open questions in physics, which concern the basic laws governing the interactions and forces among the elementary objects, the deep structure of space and time, and in particular the interrelation between quantum mechanics and general relativity.” Thanks, Wikipedia.

Around fifteen years ago, in one of my earlier professional roles, I designed the interface for the application that generated the LHC’s particle collision diagrams. These diagrams are the rendering of what’s actually happening inside the Collider during any given particle collision event and are often considered works of art unto themselves.

Designing the interface for this application was a fascinating process for me, in that I worked with Fermilab physicists to understand what the application was trying to achieve, but also how the physicists themselves would be using it. To that end, in this role,

I cut my teeth on usability testing, working with the Fermilab team to iterate and improve the interface. How they spoke and what they spoke about was like an alien language to me. And by making myself humble and working under the mindset that I was but a student, I made myself available to be a part of their world to generate that vital connection.

I also had my first ethnographic observation experience: going to the Fermilab location and observing how the physicists used the tool in their actual environment, on their actual terminals. For example, one takeaway was that due to the level of ambient light-driven contrast within the facility, the data columns ended up using white text on a dark gray background instead of black text-on-white. This enabled them to pore over reams of data during the day and ease their eye strain. And Fermilab and CERN are government entities with rigorous accessibility standards, so my knowledge in that realm also grew. The barrier-free design was another essential form of connection.

So to those core drivers of my visual problem-solving soul and ultimate fulfillment: discovery, exposure to new media, observation, human connection, and evolution. What opened the door for those values was me checking my ego before I walked through it.

An evergreen willingness to listen, learn, understand, grow, evolve, and connect yields our best work. In particular, I want to focus on the words ‘grow’ and ‘evolve’ in that statement. If we are always students of our craft, we are also continually making ourselves available to evolve. Yes, we have years of applicable design study under our belt. Or the focused lab sessions from a UX bootcamp. Or the monogrammed portfolio of our work. Or, ultimately, decades of a career behind us.

But all that said: experience does not equal “expert.”

As soon as we close our minds via an inner monologue of ‘knowing it all’ or branding ourselves a “#thoughtleader” on social media, the designer we are is our final form. The designer we can be will never exist.