The benchmarks, from our accounts
| Metric | Typical range (ZAR, 2026) | Notes |
|---|---|---|
| CPM | R60 – R180 | Higher for finance, beauty and December |
| Cost per link click | R3 – R15 | Video and UGC pull it down |
| Cost per purchase, consumables | R90 – R250 | Skincare, coffee, pet food, supplements |
| Cost per purchase, high ticket | R400 – R900 | Furniture, jewellery, equipment |
| Cost per qualified lead | R65 – R350 | Salons to solar; lower when leads land in WhatsApp |
These are the ranges across the accounts in our case studies, not platform averages. Your number will sit somewhere in them depending on three things: how well the platform can see your purchases, how good the creative is, and how much competition there is for your audience in a given month.
What a minimum budget looks like
Meta’s learning phase needs around 50 conversions a week per ad set to stabilise. At a R150 cost per purchase that is R7,500 a week, which is why we set R10,000 a month as the floor and consider R25,000 to R50,000 the range where an account can actually be optimised. Below the floor you are buying impressions, not learning.
A common mistake is splitting a small budget across many ad sets. Five ad sets at R2,000 each learn nothing; one ad set at R10,000 learns something every week.
What decides whether the ads are profitable
- Margin. A 60% gross margin on a R450 order gives you R270 to buy a customer with. A 25% margin gives you R112. The same cost per purchase is a win for one brand and a loss for the other.
- Tracking. If the pixel misses a third of purchases, Meta optimises for the wrong people and your reported cost per purchase is a third too high. Conversions API and deduplication are not optional.
- Creative. The cheapest lever. A hook that lands can halve cost per purchase in a week. Studio product shots almost never land.
- Repeat rate. If customers buy twice, the first purchase can break even and the account still prints money. That is why we run retention through email and WhatsApp alongside the ads.
The number to watch
Not CPM, not CTR, not even ROAS. Watch cost per new customer against first-order margin and expected repeat value. When that is positive and stable for two weeks, raise the budget 20%. When it turns negative, find out why before you touch the budget. That rule, applied every week, is most of what an ad agency does.
If you want to see where your account sits against these benchmarks, apply and we will look at it on the call.