Trans-inclusive Design

Late one night a few years ago, a panicked professor emailed me: “My transgender student’s legal name is showing on our online discussion board. How can I keep him from being outed to his classmates?” Short story: we couldn’t. The professor created an offline workaround with the student. Years later this problem persists not just in campus systems, but in many systems we use every day.

To anyone who’d call that an unusual situation, it’s not. We are all already designing for trans users—1 in 250 people in the US identifies as transgender or gender non-binary (based on current estimates), and the number is rising.

We are web professionals; we can do better than an offline workaround. The choices we make impact the online and offline experiences of real people who are trans, non-binary, or gender-variant—choices that can affirm or exclude, uplift or annoy, help or harm.

The rest of this article assumes you agree with the concept that trans people are human beings who deserve dignity, respect, and care. If you are seeking a primer on trans-related vocabulary and concepts, please read up and come back later.

I’m going to cover issues touching on content, images, forms, databases, IA, privacy, and AI—just enough to get you thinking about the decisions you make every day and some specific ideas to get you started.

“Tried making a Bitmoji again, but I always get disillusioned immediately by their binary gender model from literally step 1 and end up not using it. I don’t feel represented.”

Editorial note: All personal statements quoted in this article have been graciously shared with the express consent of the original authors.

How we can get things right#section2

Gender is expansively misconstrued as some interchangeable term for anatomical features. Unlike the constellation of human biological forms (our sex), gender is culturally constructed and varies depending on where you are in the world. It has its own diversity.

Asking for gender when it is not needed; limiting the gender options users can select; assuming things about users based on gender; or simply excluding folks from our designs are all ways we reify the man-woman gender binary in design decisions.

Names are fundamentally important#section3

If we do nothing else, we must get names right. Names are the difference between past and present, invalidation and affirmation, and sometimes safety and danger.

Yet, many of the systems we use and create don’t offer name flexibility.

Many programmers and designers have a few misconceptions about names, such as assuming people have one moniker that they go by all the time, despite how common it is for names to change over a lifetime. People might update them after a change in marital status, family situation, or gender, or perhaps someone is known by a nickname, westernized name, or variation on a first name.

In most locales, legally changing names is extremely difficult, extremely expensive, requires medical documentation, or is completely out of the question.

Changes to name and gender marker are even more complicated; they tend to be two separate, long-drawn-out processes. To make matters worse, laws vary from state to state within the U.S. and most only recognize two genders—man and woman—rather than allowing non-binary options.Not all trans people change their names, but for those who do, it’s a serious and significant decision that shouldn’t be sabotaged. We can design systems that protect the lives and privacy of our users, respect the fluid nature of personal identity, and act as an electronic curb cut that helps everyone in the process.

Deadnaming#section4

One need only search Twitter for “deadname app” to get an idea of how apps can leave users in the lurch. Some of the most alarming examples involve apps and sites that facilitate real-life interactions (which already involve a measure of risk for everyone).

“Lyft made it completely impossible for me to change my name on its app even when it was legally changed. I reached out to their support multiple times and attempted to delete the account and start over with no result. I was completely dependent on this service for groceries, appointments, and work, and was emotionally exhausted every single time I needed a ride. I ended up redownloading Uber – even though there was a strike against the service – which I felt awful doing. But Uber allowed me to change my name without any hoops to jump through, so for the sake of my mental health, I had to.”

Trans people are more likely to experience financial hardship, so using payment apps to ask for donations is often necessary. Some of these services may reveal private information as a matter of course, leaving them exposed and potentially at risk.

There are also ramifications when linked services rely on our data sources for name information, instigating an unpredictable cascade effect with little or no recourse to prevent the sharing of sensitive details.

These are examples of deadnaming. Deadnaming is what happens when someone’s previous or birth name is used, rather than the name the person uses now. Deadnaming is invalidating at the least, even as a faux pas, but can be psychologically devastating at the other extreme, even putting lives at risk.The experiences of trans, non-binary, or gender-variant folk can vary widely, and they live in disparate conditions throughout the world. Many are thriving and creating new and joyful ways to resist and undo gender norms, despite the common negative narrative around trans lives. Others can face hardship; trans people are more likely to be unstably housed, underemployed, underpaid, and targets of violence in and out of their homes, workplaces, and intimate relationships. The ramifications are amplified for people of color and those with disabilities, as well as those in precarious living/working situations and environments where exposure can put them in harm’s way.

Design for name changes#section5

Here’s what we can do:

Design for renaming. Emma Humphries’ talk on renaming covers this nicely. Airbnb has developed policies and procedures for users who’ve transitioned, allowing users to keep their review histories intact with amended names and/or pronouns.

Get rid of “real name” requirements. Allow people to use names they go by rather than their legal first names.

Clarify when you actually need a legal name, and only use that in conjunction with a display name field.

Have a name change process that allows users to change their names without legal documentation. (It’s likely that you have procedures for marriage-related name changes already.)

Ensure users can still change their display names when connecting with other data sources to populate users’ names.

Don’t place onerous restrictions on changes. Once someone creates a username, web address, or profile URL, allow them to change it.

Draft a code of conduct if you’re part of an online community, and make sure to include policies around deadnaming. Twitter banned deadnaming last year.

Allow people to be forgotten. When people delete their accounts for whatever reason, help them make sure that their data is not lingering in your systems or in other places online.

Update the systems users don’t see, too#section6

Identity management systems can be a mess, and name changes can reveal the failures among those systems, including hidden systems that users don’t see.

One Twitter user’s health insurance company kept their ID number between jobs but changed their gender. Another user updated their display name but got an email confirmation addressed to their legal name.

Hidden information can also undermine job opportunities:

“At a university as a student, I transitioned and changed my name and gender to be a woman. TWELVE YEARS later after being hired to work in the Libraries, the Libraries HR coordinator emailed me that I was listed as male still in the database. He changed it on my asking, but I have to wonder how long… was it a factor in my being turned down for jobs I applied to… who had seen that..?”

Emma Humphries details the hidden systems that can carry out-of-date information about users. Her tips for database design include:

• Don’t use emails as unique IDs.
• Use an invariant user ID internally, and link the user’s current email and display name to it.

Images#section7

Visuals should allow room for representation and imagination rather than a narrow subset of the usual suspects: figures who appear to be straight, cisgender, able-bodied, and white/Caucasian.

What we can do is feature a variety of gender presentations, as well as not assume someone’s gender identity if they buy certain items.

Some companies, like Wildfang and Thinx, offer a broad array of images representing different races, body sizes, and gender expressions on their websites and in their ads.

Many are also choosing not to hire models, allowing room for imagination and versatility:

“I got a catalog for a ‘classic menswear company’ that features zero photos of any person of any gender. Now if only I could afford an $800 blazer…”

Here’s what we can do:

Actively recruit diverse groups of models for photos. And pay them!

If you can’t shoot your own photos, Broadly has recently launched a trans-inclusive stock photo collection free for wide use. Avataaars allows users to create an avatar without selecting a gender.

Information architecture#section8

How we organize information is a political act and a non-neutral decision (librarians have said this for a while). This applies to gender-based classifications.

Many companies that sell consumer goods incorporate gender into their product design and marketing, no matter what. The product itself might be inherently gender-neutral (such as clothing, toys, bikes, or even wine), but these design and marketing decisions can directly impact the information architecture of websites.

Here’s what we can do:

Evaluate why any menus, categories, or tags are based on gender, and how it can be done differently:

“Nike has a ‘gender neutral’ clothing category, yet it’s listed under ‘men’ and ‘women’ in the website architecture. 🤔”

Forms#section9

Forms, surveys, and other types of data gathering are surefire ways to include or exclude people. If you ask for information you don’t need or limit the options that people can select, you risk losing them as users.

Here’s what we can do:

Critically evaluate why you are asking for personal information, including gender. Will that information be used to help someone, or sell things to your advertisers?

“Why does the @CocaCola site make me select a gender just to make a purchase? Guess my family isn’t getting personalized Coke bottles for Christmas.”

If you are asking users for their gender, you’d better have a good reason and options that include everyone. A gender field should have more than two options, or should ask for pronouns instead. When including more than binary options, actually record the selections in your databases instead of reclassifying answers as male/female/null, otherwise you risk losing trust when disingenuous design decisions become public.

Honorifics are infrequently used these days, but it takes little work to add gender-inclusive titles to a list. For English-language sites, “Mx.” can go alongside “Mr.” and “Ms.” without fuss. United Airlines debuted this option earlier this year.

Content#section10

Here’s what we can do:

Avoid inappropriately gendered language. Your style guide should include singular “they” instead of “he/she” or “s/he,” and exclude frequently used words and phrases that exclude trans folks. Resources such as this transgender style guide are a quick way to check your language and benchmark your own content guidelines.

Check assumptions about gender and biology. Not everyone who can have a period, can get pregnant, or can breastfeed identifies as women or mothers—just as not everyone who identifies as women or mothers can have periods, can get pregnant, or can breastfeed. Thinx, a company that sells period underwear, has an inclusive tagline: “For people with periods.”

Avoid reinforcing the binary. Groups of people aren’t “ladies and gentlemen” or “boys and girls.” They are folks, people, colleagues, “y’all,” or even “all y’all.”

Pronouns aren’t “preferred”—they’re just pronouns. Calling pronouns preferred suggests that they’re optional and are replacing a “true” pronoun.

Avoid reinforcing stereotypes about trans people. Not all trans people are interested in medically transitioning, or in “passing.” They also aren’t fragile or in need of a savior. Gender is separate from sexual orientation. You can’t “tell” someone is trans.

Privacy, surveillance, and nefarious AI#section11

We’ve heard the story of algorithms identifying a pregnant teen before her parents knew. What if an algorithm predicts or reveals information about your gender identity?

Inferences. Users’ genders are assumed based on their purchase/browsing history.

Recommendations. A user bought something before they transitioned and it shows up in “recommended because you bought X.”

Predictions. Users’ genders are not only inferred but used to predict something else based on characteristics of that gender. Even if you don’t tell big websites what your gender is, they assume one for you based on your interests. That kind of reductive essentialism can harm people of all genders. One of this article’s peer readers summed this up:

“Gender markers are a poor proxy for tastes. I like dresses, cute flats, and Raspberry Pis.”

Flashbacks. “On this day” algorithms remind users of the past, sometimes for better (“I’ve come so far”) or for worse (“don’t remind me”).

AI-based discrimination#section12

AI and surveillance software can also reinforce norms about what men’s and women’s bodies should look like, resulting in harrowing airline travel experiences and creating AI-based discrimination for trans people.

So, too, can trans folks’ public data be used for projects that they don’t consent to. Just because we can use AI for something—like determining gender based on a face scan—doesn’t mean we should.

Here’s what we can do:

Read up and proactively mitigate bias. AI and algorithms can reflect developers’ biases and perpetuate stereotypes about how people’s bodies should look. Use AI to challenge the gender binary rather than reinforce it. Design for privacy first. Hire more types of people who represent different lived experiences.

Toward a gender-inclusive web#section13

The ideas I’ve offered here are only starting points. How you choose to create space for trans folks is going to be up to you. I don’t have all the solutions here, and there is no singular trans experience. Also, language, definitions, and concepts change rapidly.

We shouldn’t use any of these facts as excuses to keep us from trying.

When we start to think about design impact on trans folks, the ideas we bring into question can benefit everyone. Our designs should go beyond including—they should affirm and validate. Ideally, they will also reflect organizational cultures that support diversity and inclusion.

Here’s what we can do:

Keep learning. Learn how to be a good ally. Pay trans user research participants to help validate your design assumptions. Hire trans people on your team and don’t hang them out to dry or make them do all the hard work around inclusion and equity. Make it everyone’s job to build a more just web and world for everybody.

Editorial note: All personal statements quoted in this article have been graciously shared with the express consent of the original authors.

This article is stronger and wiser thanks to Mica McPheeters at A List Apart and the following peer readers. Thank you.

Jake Atchison
Katherine Deibel, Ph.D.
Justina F. Hall
Austyn Higgs
Emma Humphries
Tara Robertson
Levi R. Walter