Rebuilding an Awwwards-winning landing page from scratch teaches things a tutorial can't. Notes from cloning Zentry with React and GSAP.
The short answerRebuilding an award-winning site teaches judgment that tutorials skip: easing, timing and trigger points. Slow the original down until you can read its durations, build the layout and scroll structure before any motion, copy the decisions rather than the pixels, make it hold up on a phone, and label it plainly as a recreation.
Key takeaways
- Rebuilding an award-winning site teaches judgment — easing, timing and trigger points — which tutorials skip because they teach APIs.
- Slow the original down in the browser's animation tools until you can read its durations and offsets, then build the layout and scroll structure before adding any motion.
- Copy the decisions rather than the pixels, make the result hold up on a phone, and label it plainly as a recreation.
Tutorials teach you APIs. Rebuilding a site that won Awwwards Site of the Day teaches you judgment — why this easing and not that one, why the animation starts at 85% viewport and not 100%, why the section pins for exactly two screen heights.
When I rebuilt Zentry's landing page, the hard part wasn't the code. It was noticing everything: the way videos crossfade behind a clip-path, the way text staggers in slightly after the layout settles, the way nothing ever moves for no reason.
Slow it down until you can read it
The fastest way to read someone else's motion is to stop it playing at full speed. Step through a screen recording frame by frame, or use the browser's animations panel to slow playback to a tenth and read off the durations, delays and easing curves.
Almost every effect turns out to be two or three simple moves layered with offset timing — a fade, a translate, a clip-path. The polish comes from the offsets, not from complexity.
Copy the decisions, not the pixels
The point of a clone isn't the finished artifact — it's the hundred small decisions you're forced to reverse-engineer. Each one is a lesson in taste from a team at the top of the field.
A useful discipline: don't open their source until your version works. The struggle is the curriculum; comparing afterwards is the exam.
Build the structure before the motion
Get the layout and the scroll structure right with no animation at all, then add motion one section at a time. On a scroll-driven page the structure is the choreography: where sections pin, how long they stay pinned, which element is driving which. If the scroll length is wrong, no easing curve will rescue it.
It also keeps the work debuggable. When a section misbehaves, you can tell whether the problem is the layout underneath or the animation on top, instead of staring at both at once.
Make it hold up on a phone
A recreation that only works at one screen width is half an exercise. The original team made their decisions for their viewport; making the same choreography survive on a phone — fewer simultaneous effects, lighter media, the same rhythm — is where you find out whether you understood why it works or only what it looks like.
Performance belongs in the same pass. Clip-path transitions, tilt cards and scroll-linked video are expensive, and a clone that stutters on a mid-range phone has copied the surface and missed the point.
Be honest about what it is
A recreation is study material, not an original work. Label it as a recreation, credit the site it is based on, and leave their assets, copy and branding out. The Zentry case study on this site says what it is in its first sentence for exactly that reason.
If you want to start one, pick a site you admire, set a deadline, and write down every decision you couldn't explain at the start. That list is the curriculum.
More writing
Elsewhere on the site
The ideas here get used rather than just written about — the projects are where the animation and performance work lives, and client work covers the production systems. If you want something built, services explains how that works, or just get in touch.