A game opened from a link has a different first minute from a game installed in advance. The player may be curious rather than committed. They may be on a phone. They may leave before a large download finishes, without ever seeing the part that took the most work.
That is the context in which I am interested in Unity's web direction. The Unity 6 launch includes tools for targeting mobile web, alongside the broader engine changes. The Unite 2024 web presentation is a useful starting point for thinking about what a browser build can become.
For a small game, the main opportunity is access. A link can put a playable idea in front of someone with very little explanation. That only works if the game respects the context in which the link is opened.
The download belongs to the design
A loading screen is often treated as the wrapper around the game. On the web, I think it is part of the first encounter. The player should know that something is happening and have a reasonable reason to wait. A blank canvas followed by a long pause spends their patience without giving anything back.
The most useful optimization may be choosing what needs to arrive first. A short opening level can establish the interaction before later assets are needed. A giant texture that looks excellent on a large monitor may contribute very little on a small phone. Audio, animation and fonts also have download costs, even when they barely affect the frame time once loaded.
Compression helps, but compressed size is not the whole story. Assets still have to be decoded and placed in memory. A page can download quickly and then stall while preparing the first scene. Measuring the route from opening the link to the first meaningful input is more informative than reporting only the build size.
Input deserves the same early attention. A mouse-oriented interface does not become a mobile interface because a finger can activate its buttons. Touch covers part of the screen. Hover may not exist. Browser gestures and page scrolling can compete with the game. A portrait layout can offer a very different usable space from a landscape desktop window.
I would design the smallest complete loop around those constraints: one understandable action, immediate feedback and a clear way to restart or continue. The loop should survive a short session. If the tab is suspended or the user switches applications, returning should not feel like entering a broken machine.
This does not mean every game belongs in a browser. Some experiences depend on large assets, precise controls or platform features that make another target more suitable. But the web is a serious place for games whose value is being easy to enter and easy to share.
The engine can help deliver the build. The designer still decides whether the build fits the moment. I would rather make the first interaction arrive promptly and feel good on a phone than demonstrate how much desktop content can be squeezed behind a link.