Gross ROAS overstated efficiency by 52%.
17.1× on gross sales, 11.2× on delivered. Judging on gross would have shown a much better account than the one actually collecting cash.
A women's fashion brand on website and messaging orders, judged on delivered revenue rather than gross sales.
Jan–Jul 2026 · website + messaging
In cash-on-delivery markets the gross number flatters. Here is what actually stuck, on EGP 327K of ad spend.
Month by month
Bar height = gross sales (delivered + returned + cancelled), labelled on top. Net ROAS = delivered revenue ÷ ad spend. Sales split from the order sheets; spend carried from the monthly media report.
What the numbers said
17.1× on gross sales, 11.2× on delivered. Judging on gross would have shown a much better account than the one actually collecting cash.
Delivery rate fell from ~70% in May to 57% in June, and net ROAS dropped from 18.9× to 9.1× on almost the same spend. The ad engine didn't change; order quality did.
Website orders delivered EGP 2.26M and messaging orders EGP 1.40M (38%). Both channels are tracked separately so each can be judged on its own.
| 2026 | Gross (EGP) | Delivered (EGP) | Delivery rate | Cancelled | Spend (EGP) | Net ROAS | Gross ROAS |
|---|---|---|---|---|---|---|---|
| Jan | 419,460 | 271,174 | 64.6% | 16.2% | 46,023 | 5.9× | 9.1× |
| Feb | 230,045 | 159,020 | 69.1% | 13.6% | 43,146 | 3.7× | 5.3× |
| Mar | 1,046,107 | 702,254 | 67.1% | 10.6% | 44,824 | 15.7× | 23.3× |
| Apr | 1,075,853 | 751,420 | 69.8% | 10.8% | 44,362 | 16.9× | 24.3× |
| May | 1,308,194 | 913,684 | 69.8% | 11.4% | 48,460 | 18.9× | 27.0× |
| Jun | 701,699 | 397,009 | 56.6% | 26.0% | 43,413 | 9.1× | 16.2× |
| Jul | 793,910 | 464,905 | 58.6% | 17.0% | 56,613 | 8.2× | 14.0× |
| Total | 5,575,268 | 3,659,466 | 65.6% | 14.2% | 326,842 | 11.2× | 17.1× |