E-commerce
Modest wear brand
Inherited from another agency that was splitting the budget across two ad accounts. We consolidated to one and concentrated spend on the winning product — return on ad spend nearly doubled and cost per purchase halved.
Case study20265.35x
Return on ad spend
2.76→5.35
5.35x
Return on ad spend
-52%
Cost per purchase
5.2x
Purchases from ads
The numbers, before and after
| Metric | 1–28 Jun 2026 · 28 days | 29 Jun – 31 Jul · 33 days | Change |
|---|---|---|---|
| Return on ad spend (ROAS) | 2.76 | 5.35 | +94% |
| Cost per purchase | EGP 518 | EGP 250 | -52% |
| Purchases from ads | 23 | 120 | 5.2x |
| Purchases per 1,000 impressions | 0.23 | 0.50 | 2.2x |
| Gross sales | EGP 48,500 | EGP 163,500 | +237% |
| Orders | 31 | 123 | +297% |
| Average daily sales | EGP 1,732 | EGP 4,955 | 2.9x |
| Cost per 1,000 impressions | EGP 121 | EGP 126 | +4% |
| Ad spend | EGP 11,916 | EGP 29,966 | +151% |
Look at the last two rows: the cost of reach barely moved (EGP 121 → 126). We didn't buy cheaper impressions — we made them convert. The same 1,000 impressions now produce double the purchases, and that comes from the offer and the creative, not from the budget.
On returns, for transparency: June's figures are the two then-active ad accounts combined. The 29 Jun – 31 Jul store window also contains a large batch of refunds that landed in the first days of July, against orders placed before we started — visible as the sharp drop at the beginning of that chart. That is why this comparison uses gross sales and ad-platform figures rather than net sales.
The proof
Direct screenshots from the store dashboard and Meta Ads Manager.





The challenge
The brand came to us from another agency running two ad accounts in parallel — a split budget, an algorithm learning from fragmented data, and two accounts bidding against each other for the same audience. June's result: EGP 11,916 in spend produced just 23 purchases at EGP 518 each, on a 2.76 ROAS. The budget itself was tight, so there was no room for scattergun testing.
What we did
- 1We shut down one of the two accounts and consolidated all spend into a single one, so the algorithm could learn from one pool of data instead of two accounts competing for the same audience.
- 2We tested every product until the winner surfaced — not a guess, an actual test with numbers behind it.
- 3With a limited budget, instead of spreading it thin we concentrated all of it on the winning product until it sold out of stock.
- 4The moment it ran out, we moved spend to the next product that had tested well — without going back to square one.
- 5For each product we produced several distinct marketing angles across designs and videos, not a single ad. Whichever angle worked got the budget.
Ready to start?
Book a free 30-minute discovery call. We'll listen, review where you stand, and tell you honestly whether we're the right fit.
Or email us at sales@scalevisionagency.com
