Stop Inventing Gestures: The Unwritten Rulebook of Spatial Interaction That Users Already Know

Somewhere right now, a spatial web developer is designing a custom pinch-rotate-flick gesture that they're absolutely convinced is intuitive. It isn't. Their users will fail to discover it, fail to execute it consistently when they do discover it, and eventually just stop using that feature entirely. The developer will never find out why.

This is the gesture graveyard. It's full of clever ideas that made perfect sense to the people who invented them and zero sense to the people who had to use them.

The spatial web has a gesture problem, and it's not primarily a technical problem. The hardware can handle complex input. The frameworks can detect it. The problem is that developers keep building interaction systems in isolation, ignoring the emerging vocabulary of spatial interaction that users are already learning from other platforms — and then wondering why their UX testing goes badly.

How Gesture Conventions Actually Form

Conventions don't form by committee. Nobody convened a meeting and decided that a two-finger spread means zoom. It emerged from iOS, got copied everywhere because it worked, and became so universal that violating it now feels like a genuine interface bug to most users.

Spatial interaction is going through the same process, just faster and messier because there are more platforms involved. Meta Quest, Apple Vision Pro, HoloLens, and WebXR on mobile all have their own interaction patterns, and they don't all agree. But there's more overlap than you might think, and the areas of overlap are where your conventions should live.

User testing data across multiple spatial platforms consistently shows a few patterns: users have higher initial success rates with gestures that map to physical world metaphors, they retain gesture knowledge better when it's consistent with other apps they use, and they're significantly more forgiving of visual design issues than they are of interaction inconsistency. An ugly button that works predictably beats a beautiful custom gesture that sometimes doesn't.

The Gestures That Have Actually Stuck

Let's talk specifics. These are the spatial interaction patterns that have achieved enough cross-platform consistency to be considered conventions at this point.

Pinch-to-select is the closest thing spatial computing has to a universal click. On virtually every hand-tracking platform, bringing thumb and index finger together to interact with an object is the primary selection mechanism. Your spatial web app should use this for primary selection. Full stop. If you're using a different gesture for selection, you're creating unnecessary cognitive load.

Gaze-plus-dwell for passive selection has emerged as the accessibility-friendly alternative to pinch selection, and it's become expected in enterprise and assistive contexts. Users look at an element, hold their gaze for 1-2 seconds, and the selection triggers. The timing has even started to converge — 1.5 seconds is the most common dwell threshold across platforms.

Palm-up to summon menus has strong adoption on hand-tracking platforms. Turning the palm upward to reveal a menu or toolbar maps to a natural "presenting" gesture that users pick up quickly without instruction. If you're building a hand-tracking interface, this is where your primary navigation menu should live.

Two-handed scaling — using both hands to resize objects — is well-established and maps to the physical intuition of stretching something. The exact gesture varies slightly (some platforms use pinch-spread with both hands, others use grip), but the two-handed nature of scaling is consistent enough to be a reliable convention.

Swipe-to-dismiss for panels and cards has carried over from mobile so thoroughly that users attempt it in spatial contexts almost automatically. If your spatial app has card-based content, swipe-to-dismiss needs to work. Users will try it regardless of whether you built it.

The Custom Gesture Failure Modes

When custom gestures fail — and they fail more often than they succeed — they usually fail in one of a few predictable ways.

Discoverability failure is the most common. Custom gestures that aren't surfaced through onboarding simply don't get used. Users don't explore interaction possibilities the way developers expect them to. They find one way to do something and stick with it. If your power-user shortcut gesture requires a tutorial to discover, most users will never find it.

Execution consistency failure happens when a gesture is theoretically possible but physically awkward enough that users can't perform it reliably. Complex multi-finger combinations, gestures that require sustained hand positions, anything that's uncomfortable to hold for more than a second — these all have high failure rates in practice, even with users who understand what they're supposed to do.

Conflict failure is underappreciated. Your custom gesture might conflict with a platform-level gesture, an OS gesture, or a gesture from another app the user has muscle memory for. When a gesture conflicts, the user experiences it as the app not responding, not as a gesture conflict. They just think it's broken.

A Decision Framework for Gesture Design

Before you design any spatial interaction, run it through this sequence.

Does a convention already exist for this action? Check what Quest, Vision Pro, and WebXR reference implementations do for the same interaction type. If they agree, use that. You're not being uncreative — you're being respectful of your users' existing knowledge.

If conventions conflict across platforms, which platform does your primary audience use? Don't try to satisfy every platform simultaneously with a single gesture. Pick the convention that matches where most of your users are coming from and document the others.

If no convention exists, does your proposed gesture map to a physical-world metaphor? Gestures that feel like doing something in the real world get learned faster and retained longer. "Grabbing" an object to move it. "Throwing" to dismiss. These work because they're already wired into how users understand physical interaction.

If you still want something novel, can it be discovered without a tutorial? This is the hardest bar to clear. Progressive disclosure through visual affordances — showing a hint when a user pauses near an interactive element — is the most reliable way to surface non-obvious gestures without requiring upfront instruction.

Is the gesture physically comfortable to perform 50 times in a session? Developers test their gestures a handful of times. Users might trigger the same gesture dozens of times in a single session. Anything that creates fatigue or requires awkward hand positions will generate complaints you won't immediately connect to the gesture design.

When Breaking Convention Pays Off

It does happen. There are spatial apps that have introduced genuinely novel interaction patterns that users learned and appreciated. But the cases where this worked share a common characteristic: the novel gesture was the core value proposition of the product, not a supporting interaction.

If your entire app is built around a new way of interacting with spatial content — if the gesture is the product — you can invest in teaching users something new. You have the onboarding budget and the user motivation. If your custom gesture is a feature in a larger app, you almost certainly don't have either.

Convention isn't the enemy of creativity in spatial design. It's the foundation that makes creativity possible. Build on what users already know, reserve your innovation budget for the interactions that genuinely require it, and your gesture graveyard will stay empty.