UX Research
Is your app urinally optimized?
Men’s washrooms and phone tab bars both settle into slots 1, 3 and 5. I tried to break that theory until only a usable layout rule was left.
Company
Self-initiated research
Role
Researcher and designer · solo
Duration
1 week · 2026
Toolkit
Literature review, falsification test, app teardown, Figma
At a glance
The problem
A pattern that sounds like a law of UX, with no evidence behind it and an obvious hole any psychologist would find in thirty seconds.
What I made
One rule that survives the evidence: never put your two most salient elements next to each other, and treat low-priority items as spacing, not filler.
8 studies, 1923 to 2013
60% chance baseline to beat
1 rule that survived
Context
There is an unwritten rule in men’s washrooms. Five urinals, and you take 1, 3 or 5. Never 2 or 4. Nobody teaches it and everybody follows it.
Then I looked at my phone. Home at one edge, Profile at the other, a big Create button in the middle, and two tabs I’d struggle to name sitting in 2 and 4. TikTok, YouTube and Strava all look like this. Same shape.
This was a self-initiated study, one week, written up as a LinkedIn carousel. The joke was the hook. The work was finding out whether anything real was underneath it.
The question
Do tab bars follow the urinal rule, and if they do, why? And is there a rule in it a designer could actually use in a critique without mentioning washrooms?
It would have been easy to stop at the pattern, give it a catchy name and post it. That version falls apart the moment someone who knows perception reads it. So success meant one thing: a rule that is still standing after I’ve tried hard to knock it down.
The rule nobody teaches and everybody follows.
How I approached it
I treated the theory as something to break, not something to prove. Four steps.
Make it precise. A row of five slots is a path of five points. If neighbours can’t both be taken, the most you can fit is three, and there is exactly one way to do it: 1, 3 and 5. Kranakis and Krizanc formalised this in 2010 as “The Urinal Problem”. So the shape is maths, not psychology.
Find the force behind the washroom. Middlemist, Knowles and Matter (1976) showed that standing closer to a stranger measurably raises stress. The washroom rule runs on repulsion.
Check whether screens have the same force. They have the opposite one. The Gestalt law of proximity (Wertheimer, 1923) says people group things that sit together, so related tabs want to be neighbours. Interfaces run on attraction.
Run a shuffle test. If repulsion really drove tab bars, the 1, 3, 5 alternation would hold in every app. It doesn’t. Apple’s Clock app (World Clock, Alarms, Stopwatch, Timers) and WhatsApp group tabs by what they do, in order, with no alternation. As soon as an app has no single dominant action, the pattern disappears.
The four-urinal check
Take one urinal away. Four urinals, two men, and the answer is 1 and 4, both ends. Taking 1 and 3 avoids standing next to anyone just as well, and nobody picks it. So the rule was never “don’t stand next to someone”. Not touching is the floor. Distance is the goal.
Urinals repel. Interfaces attract. Same edges, different rule underneath.
Making sense of it
Both rows land on 1, 3 and 5, for opposite reasons. Urinals push people apart, which is why the pattern holds no matter who walks in. Interfaces pull related things together. Add a new tab and it gets grouped with the tab it relates to, not slotted into position 2 to keep the alternation alive.
That is structural convergence, not cause and effect. Any rule that penalises neighbours on a row of five will produce 1, 3 and 5, whether it’s strangers, trees in an orchard or icons on a phone. So the useful question changed from “why do designers copy urinals?” to a better one:
What, in the way people see a screen, makes two neighbours worse than two items with space between them?
That moved the research out of the washroom and into memory, choice and vision science, where there are real studies to lean on.
What’s left once the washroom is taken out of the argument.
What I found
About a third of the original intuition survives. It’s smaller and duller than the washroom version, and it’s actually usable.
Ends get remembered. The serial position effect (Murdock, 1962): people recall the first and last items in a sequence best. That’s why Home and Profile hold the edges.
The middle gets picked. The center-stage effect (Valenzuela and Raghubir, 2009): in a horizontal row, people assume the middle option is the most important or popular. See every pricing table you’ve ever built.
Neighbours dilute each other. Visual crowding (Bouma, 1970; Pelli and Tillman, 2008): in the corner of your eye, objects closer together than about half their distance from where you’re looking blur into one. The tab bar sits in exactly that zone while you read the feed. This is the closest real thing to the urinal instinct.
Thumbs back it up. Hoober (2013) found about three quarters of phone use is thumb-driven, and the middle of the bottom bar is easy to reach with either hand.
Tab count changes the board
At five tabs you get both edges plus a centre. At four there’s no centre to win, so the edges carry everything and the two middle slots turn into filler.
The rule I’d ship
Never place your two highest-salience elements next to each other. Treat low-priority items as spacing, not filler.
A worked example: a banking app puts Home, a bright green Pay button and a badged Invest promo in slots 1, 2 and 3. Three loud things side by side, and Pay and Invest get mis-tapped. Moving Pay to the centre, Cards and Invest to 2 and 4 as quiet outline icons, and Profile to the edge gives each important action its own space. The same rule holds for pricing tables, card grids, dashboards, and a red Delete next to a blue Save.
One heuristic for nav bars, pricing tables, card grids and dashboards.
Impact
This was self-initiated, so the outcome is a tool rather than a metric: one sentence a designer can bring into a critique, backed by studies instead of a joke. It applies well beyond tab bars, anywhere two loud things are competing for the same glance.
What’s next
Audit 30 to 40 apps with five-tab bars. For each, mark the two most important tabs and the two most visually loud ones, and count how often they sit apart. Picking two of five slots at random puts them apart 60% of the time, so that’s the number to beat. I’ll publish the real hit rate even if it’s boring.
Redesign one real tab bar using the rule and test it against the original for tap speed and mis-taps.
If your tab bar has four items, look at slots 2 and 3 first.
Looking back
The catchy version was finished in an hour and would have been wrong. Trying to break my own theory took the rest of the week and is where all the value came from.
An analogy is a hook, not evidence. The washroom got people to read; the vision science is what makes the rule worth using.
Next time I’d count before writing. The audit should come first, so the conclusion follows the numbers instead of the other way round.
Next case study