Case studies / Prepd
Prepd · Meal-prep app · Saudi Arabia · Oct 2025 – Apr 2026
How Prepd cut its cost per order 68% in 60 days, and had its best month right after
A Jeddah meal-prep app that was buying growth at any price. We rebuilt the path from sign-up to first order, cut the media that was not paying back, and let the app's own order data decide the budget, through Ramadan, the category's weakest season.
Engagement closed Apr 2026 · figures from the client's Metabase and CleverTap dashboards, Nov 2025 – May 2026
4×Sales per dollar of mediaMarch 2026 vs January 2026 (Metabase sales ÷ media spend)
−68%Cost per orderCut by more than two thirds, January → March 2026
42%Sign-up completion, from 31%Started → completed sign-ups, Nov → Dec 2025 (CleverTap)
The case in 30 seconds
The challenge
Spend was at its highest and cost per order peaked with it; only three in ten started sign-ups were completed; three in four orders came from customers the app already had.
What we changed
Fixed the sign-up funnel first, cut the media that produced no orders, rebalanced the rest weekly on cost per first order, and planned Ramadan as the low season it is.
The result
Sales per dollar of media 4× in two months, cost per order down 68%, sales held at 94% of January on a quarter of the spend, and the best month in the data came right after.
Our diagnosis
The app was not short of reach. It was short of orders per dollar.
One map of the funnel and the budget before touching a campaign.
- Volume at any priceJanuary: the highest spend in the data and the worst cost per order
- Sign-up leak31% of started sign-ups completed in November
- Repeat carried the appThree in four orders from existing customers; winning a new one cost several times more
- Brand spend, no user effectThe client's own tests: awareness campaigns did not move users
- Ramadan works against lunchDaytime meal prep loses its season in February and March
The opportunity
Sign-up funnel fixCut non-paying mediaOrder-validated budgetsSeason-first plan
→Growth that pays for itselfBefore → After
What actually changed
BeforeAfter
BeforeSpend scaled ahead of efficiency
AfterBudget follows cost per order and cost per first order, week by week
Before31% of started sign-ups completed
After42% completed, on 50% more sign-up starts
BeforePlatform reports as the scoreboard
AfterMetabase orders and first orders decide what scales
BeforeEvery channel and creative kept running
AfterDaily keep / scale / stop on cost per first order; losers cut within the week
BeforeOne plan for every month
AfterRamadan planned as the low season: hold sales, protect efficiency, scale after Eid
Growth strategy
One growth system. Four connected levers.
01 — Media
Meta, TikTok and Snapchat each funded for one job, rebalanced weekly on cost per first order, not installs.
Prospecting→Sign-up→First order→Repeat
Spend fell 76% from January to March; sales fell 6%.
02 — Creative
Creator and UGC video in Saudi Arabic, judged daily: keep, scale by 25%, or stop.
Hooks that named the lunch problem beat feature ads.
03 — Funnel
Sign-up start → completion → first order, each step with its own number.
Completion went from 31% to 42% between November and December.
04 — Measurement
Metabase orders, first orders and sales judged every month; platform reports informed, the app decided.
Cost per order and sales per dollar were the two numbers on every weekly call.
Performance over time
What happened next
Metabase data, January to May 2026, indexed so the trend is visible without disclosing spend or revenue. May is the month after the engagement closed.
OrdersFirst orders
The business impact
App numbers, not platform numbers. January to March 2026 unless stated.
4×Sales per dollar of mediaJan → Mar 2026
−68%Cost per order−68% cost per order
−67%Cost per new customer−67% cost per new customer
94%Of January's sales heldOn 24% of January's media
31% → 42%Sign-up completionNov → Dec 2025, CleverTap
1,220Orders in May 2026 (after we finished)Best month in the data, 484 new customers
Media spend and revenue are confidential to the client. Figures are Metabase orders, first orders and sales ratios for the months stated; unit costs are shown, budgets are not.
Creative strategy
The creative system
The pieces that did the work, not the whole gallery.
- Format
- Creator video · TOF
Why it workedFarah's lunch-routine video: a 3.29% click-through and the cheapest sign-up applications in the account through October and November.
- Format
- Second creator · MOF
Why it workedBakhsh's cut ran beside it as the second-cheapest application source, so the account never depended on one face.
- Format
- Snapchat creator · BOF
Why it workedMageed on Snapchat produced the cheapest completed registrations of November, on a platform the app had written off.
What we learned
01
The sales did not need the January budget.
Spend fell 76% and sales held at 94%. The leak was between sign-up and first order, not in reach, so buying more reach was buying more leak.
02
Fix sign-up before buying installs.
Completion went from 31% to 42% while sign-up starts grew 50%. The same media produced twice the completed sign-ups in a month.
03
The best month came after the last invoice.
May 2026, the month after the engagement closed, was the best in the data: 1,220 orders and 484 new customers at less than a fifth of January's cost per customer. That is what a system looks like, and it is stated here with the caveat that we were no longer on the account.
Getting installs but struggling to turn them into valuable users?
We review your acquisition channels, event structure and post-install funnel, and show you where the next user should come from.
