01
Size in the fixed bar
GainedChoice and commitment in one place.
CostLess button width; long sizes must fit a chip.
Case study 01 — Flagship
Where parents in the UAE and Saudi Arabia decide what to buy, redesigned in English and Arabic on a Figma system I built from scratch.


01 — Why rebuild
The page looked older than the market it competed in, and the library under it had drifted.
| Dimension | Old FirstCry | Noon | Namshi | Mumzworld |
|---|---|---|---|---|
| Aesthetics | ✕No | ✓Yes | ✓Yes | ✕No |
| Product visuals | ✕No | ✓Yes | ✓Yes | ✓Yes |
| PDP navigation | ✕No | ✓Yes | ✕No | ✕No |
| Information structure | ✕No | ✓Yes | ✓Yes | ✕No |
| Mobile UX | ✕No | ✓Yes | ✓Yes | ✓Yes |
| Trust signals | ✕No | ✓Yes | ✕No | ✓Yes |
| Variant selection | ✕No | ✕No | ✓Yes | ✕No |
Fig. 01 — Competitor teardown with the PMs
02 — The page
Less between the shopper and the choice: the product first, the size in the button bar.


Fig. 02 — Different products, same job
03 — Three decisions
Three parts carry the purchase, so they got the most work.
A — Variant selector
Out-of-stock options stay visible, marked rather than hidden.

B — Fixed add to cart
Changing size never means scrolling back up.

C — Reviews on the page
Ratings on the page, not behind a tab, with one-tap translation.

04 — Personas
Four parents, reading the same page for different reasons.
Personas built from PM research and product data
SarahWorking mother, 32
Wants the button fixed at the bottom, and speed.
OmarFather, engineer, 35
Skips marketing for specs and the warranty.
FatimaExpecting mother, 28
Opens every section; reads reviews before the price.
AmiraStay-at-home mother, 30
Trusts other parents’ photos, and shares the link.
05 — Bilingual by design
Arabic was drawn as its own layout, not flipped at the end.


The same system carried the rest of the app.


Fig. 05 — All Categories, English and Arabic
06 — Design system
One Figma library, built from scratch in both languages, adopted by 5+ teams.
English
Nunito Four rolesArabic
Noto Sans Arabic Same roles, taller linesCore colours






07 — Adoption
Developers moved with the library, so I wrote them a guide.


08 — Four calls
Four calls, and what each one cost.
01
GainedChoice and commitment in one place.
CostLess button width; long sizes must fit a chip.
02
GainedAnswers without leaving the product.
Cost13 sheets, in two languages.
03
GainedInstalments beside the price.
CostA busier top of page.
04
GainedRTL checked screen by screen.
Cost53 English frames, 46 Arabic.
Dropped
A 60-30-10 study tried two new magentas. The system kept the brand’s own: Mulberry, #D32E9B.
09 — What it left behind
A product page in two languages, and the library later pages are built from.
MeasureAgree success metrics with the PMs at kick-off, and keep them in the file.
WebDesign desktop in the same pass, not after mobile.
ArabicFinish the Arabic set: some inputs exist only in English.
HandoffMove the guide into Dev Mode notes on each component.
10 — Companion board
The same story as a board:
Next — 02