Eight Types of Meta Sales Campaign, Ranked by What They Really Returned
By Joseph Maina · · 4 min read
Most case studies give you one big number. This one takes a single Nairobi beauty retailer's Meta account and sets its eight types of sales campaign side by side: KES 120,810 in spend, 949 purchases, all between July and October 2026. The range is wide, from 70.2x down to 15.4x, in the same account with the same products. It is about as close as you can get to a test of which kinds of offer work. The client stays unnamed. The numbers are real.
The ranking
Here is every type of sales campaign in the account. Read the chart with one caveat in mind: longer campaigns and larger budgets face a tougher test than short, small ones.
Always-on: the most revenue from the top of the table
The always-on sales campaign returned 70.2x, with KES 21,280 of spend bringing in KES 1,493,843 and 243 purchases at KES 88 each. It was the most productive campaign in the account by a distance. This suggests that consistently having something running can matter as much as any clever promotion, although it benefited from the warmth built up by everything else in the account.
Offers with a reason to act now
Next came three formats with a built-in reason to buy now. Restock announcements returned 63.4x on three one-week runs, individually between 50x and 98x. These go to an audience that already knows the product and may have been waiting for it. Product-day promotions returned 59.8x across three runs, at KES 82 per purchase, the lowest cost per purchase of any type. The strongest single run of all was one of these, at 76.8x. End-month sales returned 50.8x, which fits with when many Kenyan customers have just been paid.
These campaigns share some things: a short duration, one clear message and something real to say. "It's back in stock" is a much stronger reason to tap than "check out our products".
The strongest formats gave people a true, specific reason to buy today. None of them relied on a clever targeting setup.
Bigger events brought volume, at a lower return
The mid-year clearance sale returned 34.1x, in the lower half of the table. But it took KES 39,079, the largest budget in the group, and it brought in KES 1,331,405 and 256 purchases, the most of any type. Lower return does not make it a bad campaign. At that scale, it was one of the biggest earners. The seasonal promotion came in at a similar 33.8x, on much less.
This is why ROAS alone can mislead. A small campaign with a high multiple may have nowhere left to grow, while a big one with a lower multiple can still bring in more money. You also need to consider margins, since a clearance sale has given away some margin before the ad even runs.
The one at the bottom
The video sales campaign returned 15.4x, over 69 days, on KES 18,077, with a cost per purchase of KES 292. That is more than three times the always-on campaign's. It was the weakest, and we reduced its budget.
This does not prove that video does not work, as a return of 15.4x would be a great result in many accounts. It may have a different job, reaching colder audiences who take longer to decide. Campaign types here are not controlled experiments. But when you have to choose where the next shilling goes, this table helps you decide.
What to be careful about
These figures are as reported by Meta, on a 7-day click and 1-day view window, and usually run higher than banked sales. Campaign types differed in length, budget and season, so this is not a clean test. These numbers come from a retailer with an established brand and loyal customers. Your returns will be different, and this is not a promise of what they will be. Returns like these are unusual, so please do not treat them as a benchmark.
How to use this in your own account
You can run the same exercise in an afternoon.
- Rename your campaigns by offer type, so they can be compared.
- Export spend, purchases and revenue for at least three months.
- Group by type, and rank them by return and cost per purchase.
- Look at volume too, not only at the multiple.
- Give the winners more budget gradually, then give the weakest a new job or stop it.
- Plan next quarter's calendar around the formats that have a real reason to act.