The UI Tax of Being Everything

How I Discovered the Tax: The Audit

The product I was working on was a consumer-social app for users aged 13 to 21. The pitch was authenticity—the kind of raw, unfiltered, real product that we wanted to feel like Vine and early Musical.ly did before social media had become a performance. That was the messaging and, on paper, it was great positioning because everyone in that age group is genuinely tired of the polished version of themselves that they have to maintain on Instagram and TikTok. Our app really felt promising.

However, we realized that, in reality, the product was not delivering on the pitch. From Day 1 to Day 30, user retention was collapsing. The user base wasn’t holding steady or growing, it was actively shrinking and, while the metrics we were tracking were identifying symptoms, they were not giving us a diagnosis. We knew users were leaving, we just didn’t know why.

I needed to understand what competitors were doing that we weren’t, so I audited ten of their apps that I chose for their variety rather than their similarity. These apps included BeReal, Locket, Yubo, Poparazzi, Lapse, Snapchat, Amo, NoPlace, YikYak, and a couple of others, each in different niches, with different mechanisms and different user bases. The point of the audit was not to find a feature to copy but to figure out what the standard looked like for products that were actually retaining the audience we were trying to retain.

The pattern I noticed about halfway through the audit was that I could finish the same sentence for every product I had looked at: This is the app where ___. For example:

• BeReal was the app where the user takes one photo at the same time as everyone else.
• Locket was the app where friends’ photos appear on the home screen as a widget.
• Poparazzi was the app where friends take user’s photos.
• Lapse was the app where photos look like they came from a digital camera.
• Yubo was the app where users meet new people. They had reinforced this message so well through their TikTok marketing that the app’s identity was clear before users had even downloaded the app.
• Snapchat was the app where users send disappearing photos and maintain streaks with friends.

Every product I audited had exactly one answer. Every user interface was teaching that one answer through whatever was visible on the app’s home screen, what the user’s primary action was, what the app’s color signature was, and what items its navigation prioritized. The app’s identity was not just what the marketing claimed. It was the substrate of the user interface itself.

When I tried to finish that sentence for our app, I couldn’t. We had stories. We had locked stories that the user could unlock with coins. We had scratch cards that earned the user coins. We had a Starbucks rewards system where coins could be converted into actual coffee credits. We had the original authentic self-expression positioning. We had five different answers to the question what is this app for? A different thesis justified each of them; all of them were manifested on the home screen; and we were trying to teach all of them to new users simultaneously.

I asked ten of my friends to test the app, and the response was consistent: they said they didn’t understand the point of the app. They told me the colors bothered them. Most of them said they could just use Snapchat instead. None of them said that the features were bad or the product was broken. What they said was that they couldn’t tell what the app was for, so they defaulted to the product that had already given them a clear answer.

The reason I hadn’t seen this before the audit was that we had arrived at every feature in the product through specific reasoning, during our own meetings with its owner. We had tied scratch cards to a revenue thesis, in which users would earn coins, the coins would convert to Starbucks credits, and that loop would drive user engagement. Locked stories were part of a retention thesis, where gating premium content behind a coin economy would bring users back to the app. Authentic posting was the original product positioning the app had been built around. Each of these decisions made sense on its own. The problem was cumulative because the user interface was trying to teach three different identities at once, which meant it was teaching none of them clearly, and users were paying for that confusion through a tax on every session they didn’t come back from.

This is the sort of product blindness that no amount of internal review can fix. Each feature grew out of its own meeting, had its own roadmap, its owner, and its own metric. No meeting existed for answering these questions:

• What are we asking the user to believe this app is for?
• Is the user interface teaching that belief or contradicting it?

That meeting doesn’t exist because no one’s job description covers this. So the UI identity tax accumulates, feature by feature. While each addition is individually defensible, the cumulative effect is catastrophic.

Why Competing Identities Hurt More Than Missing Identities

The conventional way of talking about a weak product identity states that the team didn’t pick a niche. But that framing isn’t quite right and lets too many teams off the hook. Most products with identity problems have no answers to what is this app for? They have three or four answers, each of which was internally justified and has its own advocates inside the building. Each has a user interface that teaches a different mental model to a user who has not yet formed any model at all.

Competing identities are harder to fix than missing identities because removing one means removing a feature, which means killing a project that someone has built and someone else championed. That conversation is rarely about user research. It is typically about internal politics, sunk costs, and the fact that everyone who pushed for a feature has a professional incentive for it to stay. So the user interface ends up reflecting an internal compromise rather than a coherent product theory, and users pay for that compromise in cognitive effort every time they open the app.

The cognitive cost of competing identities is the part most teams underestimate. B.J. Fogg’s Behavior Model frames user actions as the product of motivation, ability, and a trigger, while the failure mode that competing identities produce is a collapse in ability. A user opening the app for the first time isn’t unmotivated; they downloaded the app, they showed up, and the trigger fired. What gets lost is users’ ability to figure out what action they are being invited to take, because the user interface is offering them five different invitations without signaling which one matters. The motivation is intact and the trigger has fired, but users cannot act because they can’t decode the product, so they leave.

The deeper reason that competing identities are so much worse than missing ones is that they aren’t just adding more for the user to process, they are forcing the user to choose between interpretations that the user interface itself cannot choose between. There is a difference between a user interface that asks users to do a lot and one that asks users to figure out what it is asking them to do. The first is heavy. The second is ambiguous. Ambiguity is qualitatively different from heaviness because no amount of additional effort resolves it. A user who is willing to put in more effort can navigate a heavy user interface. But a user who is simply willing to put in more effort cannot navigate an ambiguous user interface because resolving the ambiguity isn’t about making more effort, it is about the absence of information users would need to make sense of what they are seeing.

Think about the difference between a complicated remote control and a remote control that has five unlabeled buttons. The first is hard to use but learnable, while the second is impenetrable, no matter how patient the user is, because the meaning isn’t there to be learned. Most consumer products that have a weak identity are not complicated, they are impenetrable because essential information is missing from the user interface. The metric response a team sees in the activation funnel reflects exactly that distinction.

The resulting diagnostic question is straightforward, and you can pose it regarding your own product without conducting any new research. The question follows:

If three strangers finished their first session using an app and you asked each of them to complete the sentence this is the app that ___, would you get the same answer from all three of them or would you get three different answers? If you would get three different answers, the user interface is teaching three identities, and your retention metrics are paying the price for it.

Color Is the Fastest Identity Signal a UI Can Send

Do a recognition test in your own head. Yellow is Snapchat, the pink-purple-orange gradient is Instagram, black and white is BeReal, blue and white is Facebook, bright orange is Manzo, dark purple is Revolut, and green is Spotify. You didn’t read these user interfaces as features or color palettes, you read them as identities, and you did it in under a second. Color does all the identity work before the user has read a single label or touched a single feature.

Color is the fastest identity signal a user interface can send, because it activates before navigation, before copy, and before any feature. A user opens your app and pattern-matches its color signature against every other app they’ve ever used and, depending on what comes back from that match, your user interface either anchors on the particular app brand or it doesn’t. If a color signature is generic, the user has no anchor, and the rest of the user interface must do all the identity work alone, which it almost certainly cannot. If your color signature is distinctive, you have already established a piece of the app’s identity before the user has done anything at all.

The reason color activates faster than the rest of the user interface is that users don’t read it the way we read text or icons, they recognize it. Reading is sequential, our eyes move across letters, our brain assembles them into words, and these words trigger meaning. Color skips all of that. A pattern-match is closer to facial recognition than to reading. Users don’t decode colors, they just know them. This is what makes an indistinct color signature so expensive: it isn’t just that you have failed to provide an identity signal, it is that you have neglected to provide the only identity signal that operates before the user has done any work to understand your product. The user must actually process every other part of the identity infrastructure. Color is the one aspect of identity that lands before processing begins. Lose it and every other signal must compensate for users’ missing first impression, during every session, forever.

The reason most teams underweight the impact of color on identity is that they consider color an aesthetic decision that came late in the development process and that often gets handed to someone outside the core product team to make the app pretty. They treat color choice as a matter of taste rather than infrastructure, which is the wrong frame entirely. Color is not decoration; it is the first identity signal a user interface sends. An app sends that signal whether you have made a deliberate design choice or not. Does its color send the identity signal you intended?

This isn’t a consumer-social phenomenon either. Revolut and Monzo built their fintech identities largely through color before either had any distinctive features because in a category where every product has the same core function—move money, view your balance, and send a card—color was one of the few signals that could carry the message about the bank’s identity before the user had explored anything. Notes apps, browsers, productivity tools, all lean on color as a primary carrier of identity because, in crowded categories with similar features, color is often the only piece of identity infrastructure that scales to a glance.

When we redesigned the app I was working on, the color change was one of the things I had held the firmest line on. The pre-redesign palette was dark in a bad way, with colors mixed up in a way that strained the eyes and signaled nothing in particular about what the product was. The post-redesign palette had a white base with a rainbow gradient running through purple, pink, and turquoise, with yellow accents. The point wasn’t that the new colors were more beautiful, the point was that they were distinctive, and distinctiveness is what color-identity infrastructure provides when it is working.

However, the honest caveat is that color alone cannot make an identity. In fact, distinctive colors sitting on top of a product that doesn’t yet know what it is for are decoration, not identity, because there is no underlying answer on which the colors can anchor. Color is part of the identity infrastructure, not all of it, and it must work alongside a coherent answer to the question what is this for? or it is just a nicer-looking version that causes the same confusion.

The diagnostic question for color is the inverse of the recognition test, as follows: If your app’s color signature appeared on screen without any logo, copy, or recognizable user interface, would users who have seen your product before recognize it? If not, you have a color-identity gap that is one of the cheapest pieces of identity infrastructure to fix.

Gamification Follows Identity but Doesn’t Substitute for It

The other thing that becomes obvious once you start looking at consumer products through the identity lens is that gamification mechanics do retention work only when they are reinforcing an identity that already exists. Snap streaks work because people already know that Snapchat is the app for lightweight, continuous contact with close friends, and a streak is literally the mechanical expression of that promise. Duolingo's streak works because Duolingo is already the app for daily language practice, and a streak is the mechanical expression of consistent practice. Strava’s streak works for the same reason.

Take the same identity mechanism and transplant it into a product that hasn’t decided what it is for, and the mechanism doesn’t carry any identity meaning. It just becomes a chore users are being asked to commit to before they have decided whether the product is worth committing to. While gamification mechanics are identity amplifiers, they are not identity creators. So trying to use them as a substitute for identity is one of the most expensive mistakes a consumer-product team can make because it adds engineering and design cost without addressing the actual problem.

This approach also gives a third diagnostic, which I think is the sharpest of the three, as follows: If you took each of your gamification mechanisms and transplanted it into a different app in a different category, would they still mean something to users? If you could lift a streak, a coin economy, a scratch card, or a badge out of your product and drop it into a completely different product without losing its meaning, the gamification mechanism isn’t doing identity work in your product. It is running interference for the absence of identity, and your retention numbers would reflect that.

In the case I’m describing, the gamification stack ran on a different track from the product positioning. Scratch cards earned coins, coins unlocked locked stories, coins also converted to Starbucks credits, and none of these mechanisms encoded anything about the product’s stated identity of authentic self-expression. Each mechanism made sense based on its own thesis, but as a stack, they were running interference for the absence of a coherent answer to the question what is this app for? Users felt the disconnect even when they couldn’t articulate it, and the retention curve told us that they had.

What UX Design Could Fix and What It Can’t

The redesign work that came out of the audit moved real metrics. We reduced screen depth and tap counts, and simplified navigation so the home screen signaled fewer competing answers to the question what is this app for? We also established a distinctive color identity and pruned back the parts of the gamification stack that were most disconnected from the product’s positioning. From Day 1 to Day 30, retention improved by more than twenty percent, and the improvement held over the months that followed. The product is still active today and has a substantial user base that is still growing. The UX design work we did was not cosmetic, and the metrics responded in the way that you would expect when a user interface gets clearer about what it is for.

But the strategic question of which identity the product was actually for did not get fully resolved at the design layer because it couldn’t. Authentic self-expression, scratch-cards-for-rewards, and coin-locked content were still all present in the product in some form, and a UX redesign could clean up how we presented these competing layers to a new user without resolving which was the actual answer. That resolution requires product-strategy work, not user interface work, and it sits outside what any redesign can do alone. The retention lift came from making users’ cumulative confusion less expensive for them to navigate, not from removing the underlying conflict between the three identities the product was trying to carry.

This is the part of the story that I think matters most for anyone reading this article and trying to apply this framework to their own product. UX design can reduce the cost that users pay for an unresolved identity, but cannot itself resolve the identity. Once you’ve cleaned up a user interface and user retention has improved but plateaued, the remaining gap is not necessarily a UX design problem. It could be a product-strategy problem that is leaking through the user interface. The most useful thing that UX design can do at that point is to name the leak clearly enough that the Strategy team is forced to address it. The alternative is another redesign cycle that polishes the symptoms while the root cause continues to charge users a tax they cannot afford to pay.

The Identity Tax Audit

I’ve compressed this overall audit process into an identity-tax protocol that you can run on your own product, as follows:

1. The sentence test—Ask five people who are not on your team to use your app for a single session, then ask each of them to complete the sentence This is the app that ___. If you get five different answers, your user interface is teaching five identities, and the variance you are seeing is the cost of that confusion showing up in your activation funnel before it shows up in your retention curve.
2. The color signature test—If your app’s color palette appeared on screen with no logo, no copy, and no recognizable user interface, would users who have seen your product before recognize it? Color is the fastest identity signal that a user interface sends, and an indistinct color signature is one of the cheapest pieces of identity infrastructure to fix once you’ve recognized that it is infrastructure rather than aesthetic.
3. The transplant test—Could you lift each of your engagement mechanisms out of your product and drop it into a different app in a different category without losing its meaning? If yes, the mechanism isn’t doing identity work, it is masking the absence of identity, and the retention lift you’re hoping it will produce won’t materialize until you’ve answered the underlying identity question.
4. The competing identities test—List every feature on your home screen, then group those features by whatever answer they imply for the question What is this app for? If you end up with more than one group, the user interface is teaching competing identities, and users are paying the cost of this competition for every session.

The output of this protocol is not a redesign brief, it is a diagnosis that tells you what the user interface is teaching, what the marketing is claiming, and where the gap between the two sits. What the team does about the gap is a strategy conversation, not a design conversation. Conflating the two is how teams end up with a beautifully redesigned user interface that still doesn’t retain users because the redesign addressed symptoms while the strategic conflict that produced them remained unresolved.

Every UI Answers the Question: What Is This For?

The job of a consumer-product user interface is not just to be usable, it is to teach users what your product is for, through what they can see and what they can do. This teaching will happen regardless of whether you’ve deliberately decided what the lesson should be or lessons have accumulated because every element of the user interface—the home screen, the navigation, the color, the gamification, and the empty states—is signaling something to the user who has not yet formed any model of your product. If your team has agreed on what the lesson should be, the user interface can teach it coherently. However, if the team has not agreed, the user interface will still teach a lesson, but it will be confused, and every user in every session will pay the cost of that confusion for as long as the confusion persists.

This is irreducibly a UX design problem, even though most product organizations have filed such problems somewhere else. Positioning, marketing, brand, and strategy all contribute to the answer at which the user will arrive, but it is the user interface that actually delivers the answer, and the user interface is the only place where the answer becomes real for the user. Therefore, the people who are closest to the user interface have a responsibility that most product cultures don’t yet recognize: to look at what the user interface is actually teaching to users and to compare it honestly against what the team thinks the user interface is teaching. The gap between these two is where the identity tax lives.

Every user interface is answering the question What is this for? But is the answer the one you intended?