Ideas in.
Products out.

They already track. They lack proof.

Strength logging is a crowded shelf, and almost every app on it runs on the same engine: streaks, badges, a feed. Before designing anything I sat down with four serious lifters. The same three signals kept coming up.

I lift too, which helped me hear the thing nobody said out loud: a serious lifter already knows whether they showed up. So the bet: build for people who already train, and let the casual market keep its streaks.

Two decisions, locked early.

What they were missing was proof that the work is working. Two decisions locked here. Reps-in-reserve sits at the centre of every logged set, and strength over time is shown honestly, so an easy day never gets dressed up as progress.

Show progress without lying.

The open question was how to show progress without lying about it: an honest max estimate adjusted for effort, stall detection, muscle balance, personal records by rep range. Once early users were in the private beta, a feature survey ranked what got built.

I also built a streak counter, liked it, and killed it. A streak is the exact game Notch exists to end.

Ship by subtraction.

Converging meant cutting. The exercise library shrank to the lifts a serious program actually uses, and features I was excited about moved to after launch. I designed every screen and directed AI dev tools through the build, so the design lives in the code. The launch metric is retention.

• doneOBSIDIAN system and every screen — built and working.
• doneBrand and waitlist live at notchlift.app.
• doneFull QA pass — green test suite, a security review, a class of crashes designed out.
• nowPrivate beta with real lifters, last pre-launch fixes.
• nextApp Store submission and the 2026 launch.

Cold traffic hit a wall of cards.

The homepage was built for people who already knew the brand. It opened on sixty-plus challenge cards in one dense grid: no pitch, no order, nothing telling a first-timer where to look. Analytics showed them arriving cold, meeting the wall, and leaving without picking one.

scroll ↕

Not a catalogue. A sequence.

The site had plenty of challenges and no path through them. So the brief became sequencing: lead with the pitch, feature a handful of challenges, put the proof near the top, and give each section room to be read at a glance.

Pitch, then path, then proof.

With a developer and a CRO I rebuilt the page in that order: pitch up top, a five-step how-it-works, the proof, then a short list of popular and licensed challenges. The full catalogue stayed one click away. Underneath it, a shadcn design system so design, dev and campaign teams work from one source.

scroll ↕

The test settled it.

Organic click-through rose 68% against the old homepage. Most of the lift came from moving the catalogue one click back and opening with a pitch and proof instead.

Three tables, three questions at once.

Three plans in three separate tables gave the page three competing calls to action. Hesitation costs the most at the money step, and this layout asked people to solve a comparison puzzle right there.

Which plan, then one button.

Three tables side by side turn picking a plan into comparing three plans first. The brief narrowed to a single sentence: make the page ask less.

Narrowed live, round by round.

I compressed the three tables into one plan selector with a single button underneath: the same comparison, one control. Four to five live A/B rounds with a developer and a CRO, each version tested against the last.

Pages stopped competing with themselves.

The final round settled at +11% conversion, real money on a small bar. Same fix as the homepage, one step deeper in the funnel.