Game UX

Practicality is the unglamorous layer: is the action physically available to this player, on this input, at this distance from the screen? Think of a controller player navigating a mouse-shaped menu from ten feet away, a colorblind player reading a rarity system that’s pure hue, or a new player whose one needed button hides behind a hold-input they don’t know exists. If the answer to “can they do it” is no, nothing else on the screen matters.

Usability is the second layer: do they understand what just happened, and could they predict what happens next? This is where hierarchy, feedback, and consistency live. It’s the craft most of the discipline’s literature already covers (and covers well).

Feeling is the third layer: does the screen have weight, cadence, intention? It’s real and it’s load-bearing, enough so that this medium earns its own principle for it below. One carried rule from my draft book lives here: make it look intentional. Players will forgive a plain screen, but they won’t forgive a confused one, and they read accidental-looking design as a statement about how much you cared.

Norman was right, and games prove him more right than anywhere else: feeling does real work in a game, the kind no garnish can do. But the hierarchy still runs one direction. I’ve watched teams budget the feeling layer first (gorgeous motion studies for a menu whose information architecture didn’t exist yet) and the polish always ends up decorating a hole. Build the floor, then the walls, and hang the art last.

The best session-layer case study of the era isn’t mine to claim, but I spent five years close to it. Apex Legends’ ping system was designed as squad coordination for strangers: three players, no shared language, no working mic, forty seconds to agree on a plan. Respawn built it honest, with roughly a month of playtests with mics muted and fake names, so developers couldn’t lean on knowing each other (Game Developer, 2019). It launched with the game in February 2019, and PC Gamer called it “a tiny miracle” within days.

Then the dividend arrived: deaf and hard-of-hearing players competing on equal footing, players with social anxiety opting out of voice chat without opting out of teamwork, non-native speakers coordinating across languages they don’t share. None of that was the brief, but all of it was the result. EA patented the system, then pledged it royalty-free in its 2021 accessibility patent pledge, and some version of it now ships in nearly every multiplayer game.

Hamilton’s framing is the sharpest in the field: challenge is the product, so the work is triage. Keep the barriers you designed and kill the ones you didn’t. The ping system never touched Apex’s designed challenge; it removed an accidental barrier, the assumption that coordination requires voice. That’s why it worked, and that’s the transferable lesson. When you design for the player in the hardest position (no mic, no shared language, no hearing) you aren’t making the game easier, you’re making it reachable. The constraint isn’t a tax on the design, it is the design.

Instrumented playtesting is one of the great gifts games gave back to UX, and I use it without apology. Telemetry is unbeatable at finding failure modes: the funnel cliff where new players evaporate, the screen where response time drifts past the Doherty threshold and the interface stops feeling like a conversation, the juice curve from the last chapter bending back down at the extreme. When players get lost, the data will tell you where, with merciless precision.

What the data can’t do is author. Take the yellow-paint argument that boiled over in 2023–24: the playtest record says lost players churn and explicit markers fix it. That’s true, but it’s a floor. “Paint the climbable surfaces yellow” is not a target; it’s the cheapest patch on a missing language. Wayfinding is light, color, composition, and landmark, and paint is the last layer. The grown-up answer is to make its intensity a player option instead of a default insult.

Hodent built much of this discipline, and I’d sharpen exactly one word: best has to be authored, never simply found. Data can identify the worst experience in your game with total reliability; it can’t specify the best one, because the best one doesn’t exist yet. Teams that let metrics set the targets converge on the same local maximum as everyone else’s metrics (which is why so many live-service menus feel like the same menu). Measure the floor, but author the ceiling yourself.

My lineage taught me task-completion UX: reduce time-on-task, kill steps, get the user to done. On the web that ethic is correct, because the user has somewhere better to be. A game is the somewhere better. The player opened your product to spend an evening inside it, and that flips the math on a whole class of decisions.

Ceremony becomes an asset: the lock-in, the slow reveal of a reward, the drop ship door opening. On the web these are dark-pattern-adjacent delays; in a game they’re how a session gets a shape, with anticipation, ritual, and punctuation. Menus have game feel too. The cadence of navigation, the weight of a confirm, and the sound of a tab change all draw from the same feel budget as the guns. That’s chapter 01’s argument, continued at a longer timescale.

The discipline is knowing which screens are ceremonies and which are errands. Rejoining a match, recovering from an error, finding a setting? Those are errands, so apply the full web ethic and make them vanish. The moments a player will remember from tonight are ceremonies, so give them weight on purpose. Both failure modes are real: ceremony on an errand is friction the player never asked for, and errand-efficiency on a ceremony spends a session beat for nothing. And to name the obvious abuse, “players want to stay” describes the contract, not a license to trap anyone. Friction bolted onto a cancel flow is just a dark pattern wearing ceremony’s clothes.

Two carried rules from my draft book frame this section. First: no one reads your UI text. Players parse shape, position, color, and motion, in roughly that order, with words a distant last (your beautiful tooltip is a speed bump they’ll vault). Second: be consistent, or at least predictable. Perfect consistency is unaffordable in a live game; predictability is non-negotiable. From those two rules fall the five tools I reach for on every screen.

Language direction. When you do spend words, they all pull one direction. One name per concept, decided once, enforced forever: if it’s a Legend on one screen it isn’t a Character on another, and the currency in the store is never a nickname in a tooltip. Synonyms read as new systems, and players will go hunting for a difference that doesn’t exist. Naming is cheap until it ships; after that, every rename is a patch you’re applying to player memory.

What belongs to what. Ownership must be visible before it’s readable: the price married to the thing it buys, the stat to the weapon it describes, the timer to the thing that expires. Proximity, containment, and alignment do this work; explanatory text does not (see rule one). In my experience, most feedback that says “this menu is confusing” is ownership ambiguity wearing a trench coat.

Nav nouns and button verbs. Navigation goes to places, so name them with nouns: Store, Loadout, Ranked. Buttons do things, so name them with verbs: Equip, Buy, Ready Up. The question a player asks before every press is “does this take me somewhere, or do something?” and grammar answers it faster than reading does. A button labeled with a noun is a coin flip when you’re in the player’s shoes.

Smart defaults. The default is the most-used feature you will ever ship, because most players never open the menu where you let them change it. The default loadout, the default sort, the pre-selected option: each should be the choice you’d make for a smart friend on their first night. Defaults are also where this site’s thesis bites at the smallest scale. Every default is a statement about who you think the player is, and lazy ones get heard.

Expand over time, don’t remove. Sessions accrue muscle memory, and muscle memory is a save file you don’t get to delete. Add surface as players grow: reveal tabs, deepen screens, unlock density. But removing something players have learned charges them a relearning tax and breaks something subtler, the trust that learning your game is a safe investment. Reorganize when you must, but migrate players over instead of amputating what they know.