The problem
EZI sells nano engine treatments on o2proshop.com. A driver doesn't search for a product name; they search for what their car is doing wrong.
What nobody noticed
The client's PDF brief asked for a problem router, not a catalogue: the visitor picks the symptom (weak pull, higher fuel use, an oil change due, a diesel engine) and lands on the product that fixes it.
What I built
- A real 3D can rendered in raw WebGL2: 1,152 triangles, one draw call, drag to spin.
- A symptom-based router that sends each visitor to the product that fits.
- Arabic, right to left, mobile-first; the scroll is never hijacked.
- Rendering stops once the can leaves the screen; devices without WebGL2 get the flat packshot.
- About 64 KB in total, with no framework.
My role in the projectA 3D product page built from the client's PDF brief: design and build.
Screens
HeroThe can and the question
Problem routerPick the symptom, get the product
Red or blueChosen by the car's mileage
The numbers
-
The file that spins the can
459 KB →47 KB
How it was measuredFile sizes on disk. A baked 30-frame turntable covered only ±20° before the seam showed; one unwrapped label texture covers the full 360°.
Honest scope
- Client brief; the page is shown here as captures because the demo is offline.
- The product, its name and the label artwork belong to EZI and O2PRO.
- No sales or conversion figures are claimed.
Demo offline
All workSend your
store link.
I'll reply with a 1-minute video and three things nobody told you.