DEV Community

Cover image for Eight Things That Surprise Everyone on Their First VR Project
Alina Volkava
Alina Volkava

Posted on

Eight Things That Surprise Everyone on Their First VR Project

Building for a headset looks like building for any other platform right up until the moment it doesn't. The tooling is familiar. The engines are the ones you might already know. The mental model you bring from web or mobile work carries you through about the first two weeks.

Then you hit the things nobody warned you about, and they're all in roughly the same place: the assumption that a screen is a screen.
Here's what tends to catch people out, gathered from the kind of retrospectives where somebody says "if we'd known that in month one."

1. Performance isn't an optimization phase, it's a constraint you design around

On a website, a slow frame is an annoyance. In a headset, it's a physical symptom.

The rendering budget is unforgiving and it's non-negotiable in a way most developers have never experienced. You're drawing two viewpoints, at high refresh, with no tolerance for the occasional hitch. Miss the target and users don't file a bug report, they take the headset off and feel unwell for twenty minutes.

The practical consequence: performance can't be something you address before launch. Every asset decision, every environment choice, every interaction has to be made inside a budget you established at the start. Teams that treat it as a late-stage cleanup task end up cutting the features that justified the project.

2. Comfort is a hard requirement, not a quality bar

Related, but distinct enough to state separately.

Motion sickness in VR isn't a preference issue like disliking a color scheme. Get locomotion wrong and a meaningful share of your users physically cannot use what you built. It doesn't matter how good the content is.

There's an established body of practice here: how movement should work, what camera behavior to avoid, why you never take control of someone's view, how acceleration should be handled. None of it is intuitive if you're coming from flat-screen work, where dramatic camera moves are a tool rather than a hazard.

The uncomfortable part is that the developers building it are usually the least susceptible people in the room. You acclimate. Something that feels fine to the team after three weeks of daily use can be unusable for a first-timer, which means your internal testing systematically misleads you.

3. There are no interaction conventions to fall back on

On the web, you inherit decades of accumulated agreement. Users know what a button looks like, that underlined text is a link, that a form submits. You don't design those, you use them.

In VR, very little of that exists. How does someone pick something up? Grab and hold, or grab and toggle? What does an unreachable object do? How do you show that an action is available? How does a menu appear without breaking presence?

Every one of these is a decision, most of them have several defensible answers, and users arrive with wildly different prior experience. Someone who's played a lot of VR games expects one thing. Someone who has never worn a headset expects nothing at all and needs to be taught, inside the experience, without a tutorial that feels like homework.

This is design work, and it's usually underestimated by roughly a factor of three.

4. Testing requires bodies in rooms

You cannot evaluate a VR experience by watching a recording. You can watch what someone saw, but not how they moved, where they hesitated, that they instinctively reached with the wrong hand, or that they kept turning toward a wall because a sound cue was mispositioned.

That means physical test sessions with real users, which is slower and more logistically annoying than any remote testing you've done. It means scheduling. It means space. It means watching people's actual bodies for signs of discomfort they won't mention.

It also means your testers get worse over time as they learn the environment, so you need a supply of fresh ones. Budget for this explicitly, because it's the step most likely to be skipped under deadline pressure and the one whose absence shows up most clearly in production.

5. The 3D content pipeline is its own discipline

The software is often not the long pole. The environments, models, textures, and audio are.

Someone has to produce or source all of it, at a quality level appropriate to the use case, within the performance budget from point one. Assets that look great in a modeling tool routinely need substantial rework before they can ship in real time. Scanning a real space produces data that needs cleanup. Getting an accurate representation of a client's actual facility means either site capture or working from drawings, and both are slow.

Teams staffed entirely with engineers discover this in week four. An experienced VR app development company plans the content pipeline alongside the software from the start, because the two schedules are coupled and the art side is frequently the one that determines the launch date.

6. Device fragmentation is real and it's about capability, not just resolution

Standalone headsets, tethered ones, different controller schemes, different tracking capabilities, different input assumptions, wildly different processing power. A build that runs comfortably on a workstation-tethered device may be impossible on a mobile chipset without significant rework.

Deciding your target hardware early is one of the highest-leverage choices in the project, and it should be driven by deployment reality rather than by what the team happens to have on the desk. If the eventual users are field technicians at twelve sites, they're getting standalone devices, and everything about the build has to accept that constraint from day one.

Retrofitting for a lower-capability target is not a port. It's frequently a rebuild of the visual layer.

7. Onboarding consumes more of the experience than you'd guess

For a specialist audience, you can assume some familiarity. For everyone else, assume none.

A meaningful portion of users have never worn a headset. They don't know how to adjust the strap for a clear image, so they'll conclude your visuals are blurry. They don't know the controllers, don't know they can turn around, don't realize they can move closer to something.

All of that has to be handled inside the first ninety seconds, gently, without a wall of instructional text that nobody reads. Getting this right is a substantial design task, and getting it wrong means users form an opinion about your application based on a hardware fitting problem.

8. Nothing about it is a one-and-done delivery

Platform runtimes update. Store requirements change. Device generations arrive and older ones fall out of support. The procedure your simulation teaches gets revised. Fleet management, build distribution across sites, and version consistency become ongoing operational work.

The projects that survive treat this as a maintenance line in the budget from the start. The ones that don't produce something that works beautifully at launch and is broken in a way nobody notices for four months.

The through-line

Almost every item here comes from the same root: in VR, the user has a body, and their body is inside the thing you built.

That single fact is why performance becomes a health issue, why testing has to be physical, why interaction can't be borrowed from flat interfaces, and why onboarding matters so much more than it does on a screen you can glance away from.

Bring that assumption in on day one and most of these surprises stop being surprises. Bring in a mobile-app mental model and you'll rediscover all eight, in order, at the worst possible moments.

Top comments (0)