Your account manager suggests Advantage+. You spend £8,000 a month on Meta. Manual campaigns return 3. Advantage+ reports 4.5. That looks like a 50% improvement for no extra work.
Advantage+ lets Meta decide who sees your ads. The 4.5 counts every sale it touched, including people already coming back. So the question is how much of the 4.5 the ads caused.
What the 4.5 hides
Advantage+ finds the cheapest path to a purchase. Often that is your warm audience: past buyers, recent visitors, people about to search your name. They see an ad, they buy, it counts as a win.
Say a holdout test (below) shows 60% of credited sales were caused by the ads:
£21,600 on £8,000 is a real return of 2.7, not 4.5. Move the slider to see how the answer changes with your own share.
Against a break-even of 2.5, 2.7 is thin. It is the same trap as platform ROAS, larger because Meta picks the audience.
The rule
Judge Advantage+ on total sales against total spend, from a test. Not on the ROAS in Ads Manager.
Step 1: check you are ready
- Creative: 15 or more active ads across video, static and catalogue. With 5 similar product shots, all of them go to the same warm group.
- Purchases: a steady weekly flow. Meta’s guide asks for about 50 a week to leave learning, but some accounts settle much earlier and some never settle.
- Budget: enough that learning does not eat most of it. We use £5,000 a month as a rough floor.
Fix any gap before you switch.
Step 2: run a holdout
Split your UK regions into matched pairs by population, urban and rural mix, and past order volume. Switch Advantage+ on in one set and off in the other for 4 weeks. Compare total sales divided by total ad spend in each.
A gap clearly bigger than the normal week-to-week swing between the 2 sets means Advantage+ adds sales. A small or negative gap means it collects demand that was already there.
A partial result is common. Do not switch it off. Size the budget to what the test showed, and cover finding new people with a manual campaign.
Step 3: run it beside manual campaigns
A starting split of your Meta spend:
- Advantage+ sales, 50 to 60%. Broad finding, large creative library.
- Manual prospecting, 25 to 30%. You see results by audience and brief creative from them.
- Manual retargeting, 15 to 20%. Cart abandoners, high-value viewers, lapsed customers, with message order and frequency in your hands.
When to leave it off
- New account, under about 6 months. Too little purchase history.
- Product launch. No sales history to match against. Lookalikes of your best customers give a clearer start.
- Under about £5,000 a month, our rough floor. Deliberate audiences are more efficient at that size.
Related reading
- Why your Meta ads stop working faster than they used to
- Platform ROAS is a lie
- What is blended ROAS?
Not sure how much of your Advantage+ result is real? Book a free call and we’ll read it with you. Or see how our Meta ads team plans the split.
Quick answers
What is Meta Advantage+ sales (formerly Advantage+ Shopping)?
Advantage+ sales campaigns are Meta's automated campaign type. You set a budget and a conversion goal. Meta decides who sees the ads, where, and which of your creatives to use. It can mix new audiences and past visitors in one campaign.
Do Advantage+ campaigns beat manual campaigns?
Sometimes. It tends to work when you have plenty of creative, a steady flow of purchases and enough budget. With little creative or a small budget it often leans on people who already know you, which flatters ROAS and adds little new revenue.
How do I know if Advantage+ is working or just claiming credit?
Run a regional holdout. Switch it on in some matched regions and off in others for 4 weeks, then compare total sales against total ad spend in each. If the Advantage+ regions are clearly ahead, it is adding sales. If not, it is collecting sales that would have come anyway.
Should I run Advantage+ alongside manual campaigns?
Most accounts should. Advantage+ for broad finding, manual campaigns for tight retargeting and for products where you need control. Running only Advantage+ means you cannot see what the algorithm does with your budget.
