There is an appealing temptation with VR: if the screen can surround the viewer, the experience should become enormous. An entire city, a complicated control scheme, a world with something happening in every direction. Mobile VR suggests a more useful starting point. Choose one experience that deserves a person's attention, then make entering and using it pleasantly simple.
Google's Daydream announcement describes a platform spanning compatible phones, a headset and controller, and a software experience. In May 2016 this is an announced direction, not a reason to assume every phone suddenly supports the same VR workload.
The budget is continuous
A conventional application can sometimes conceal expensive work behind a loading state. In VR, the view keeps responding to the person wearing the display. Rendering cost is connected to comfort because the world must remain responsive while the head moves.
A small scene can still be expensive. Transparent layers, large textures and repeated work can cost more than the apparent number of objects suggests. A beautiful still image tells us little about whether the experience remains stable during movement or after a longer session on a warm phone.
That makes scope a technical tool. A single room offers opportunities to control what is visible, reuse resources and avoid spending time on distant detail. It can also be a strong creative decision. An intimate scene does not need to apologize for lacking a horizon full of buildings.
Input should fit the action
A controller offers a more explicit way to point and act than waiting for every gaze target to activate. It does not mean a desktop interaction should be copied into the headset. A tiny toolbar remains tiny. Dragging through nested menus remains tedious, even if the pointer floats in space.
Imagine a scene about arranging three objects to change a sound. The important actions are selecting an object, moving it within a constrained area, and hearing the result. A short visual cue can explain the relationship. There is little reason to add an inventory, a character controller or a settings wall before that relationship feels good.
Constraints can also protect the visitor from uncertainty. If an object can only move along a rail, the interface can show the rail. If a scene is intended to be viewed from one location, its composition can work from that location instead of implying freedom it cannot support.
Leave room for the person
I would spend some of the saved complexity on the beginning and end. What does the visitor see before putting on the headset? Is the first useful interaction obvious? Can they stop without hunting through the scene? Those questions are less glamorous than adding another environment, but they shape the experience more directly.
The purpose of cutting features is to make the remaining actions legible. A small virtual place that responds consistently can be more absorbing than a large one that repeatedly reminds you how much work the phone is doing.