Where Advantage+ shopping campaigns fit and where to keep manual control
Meta began rolling out Advantage+ shopping campaigns on August 15. Test them against your best manual campaign, and keep manual control where audience rules and creative learning matter.
Meta announced Advantage+ shopping campaigns on August 10 and began rolling them out to ecommerce and retail advertisers worldwide on August 15. Meta describes them as an AI-driven campaign type that removes manual steps from ad creation and automates up to 150 creative combinations at once. The point for the buyer is simple: you hand over more decisions to Meta's system and, if Meta is right, you get lower costs for it.
What Meta has claimed, and what it has not shown
Meta's figure is that in a study of 15 A/B tests, Advantage+ shopping campaigns had a 12 percent lower cost per purchase conversion than advertisers' business-as-usual ads. That number comes from Meta, and the announcement does not describe the test length, the types of advertisers or the variance. It is a reason to run a test, not a forecast for your account.
The announcement also does not describe, in detail, which settings advertisers can still control. Before you plan around the product, open the campaign setup in your own Ads Manager and write down what is fixed and what is adjustable. That list decides where the product can sit in your account.
Test it like a budget decision, not a feature
The cleanest way to judge an automated campaign is to give it a defined slice of budget and compare it with your best-performing manual campaign over the same period. A few rules keep that comparison honest:
- Same product catalog, same goal. Both campaigns should optimize for purchase and draw on the same products and offers.
- Same creative pool, to start. If the automated campaign gets fresh creative and the manual one gets stale ads, you are testing creative, not automation.
- A fixed budget share. Something like 10 to 20 percent of prospecting spend is a reasonable starting point for a hypothetical store spending $50,000 a month, so $5,000 to $10,000. Adjust to what your volume supports; it needs enough purchases per week to leave the learning phase.
- Judge on your own numbers. Compare cost per purchase and return on ad spend using your backend orders, not only Meta's reporting, since both campaigns share the same attribution limits.
- Run it long enough. Give it several weeks and avoid mid-test edits that reset learning.
Watch for new-customer inflation
An automated campaign that combines prospecting and remarketing can look excellent because it finds people who were about to buy anyway. Look at the share of purchases from people who are new to your store, using your own customer data. If the campaign's returns are driven by existing customers, it is performing as a remarketing campaign, and you should compare it against your remarketing, not your prospecting.
Where the automated campaign likely fits
- A broad catalog with a steady purchase volume and a pixel that reports purchases reliably.
- Accounts that spread budget across many small ad sets that never leave learning. Consolidating into one campaign can help them.
- Teams that can supply a good range of creative assets but lack time to manage dozens of ad sets.
Where to keep manual control
- Creative learning. If you want to know why an angle works, a structured manual test with one variable changed isolates it. A system that mixes combinations gives you results without much explanation.
- Audience boundaries. If you need to exclude certain groups, serve different offers to different segments or keep prospecting strictly separate from remarketing, you need the control a manual setup provides.
- Launches and promotions. A new product with little history, or a time-boxed sale with a specific message, is easier to steer by hand.
- Low volume. With few weekly purchases, any campaign type struggles to learn. Fix the volume problem first.
- Reporting needs. If your finance or retail team needs spend broken out by audience or placement, confirm that the automated campaign can report that before you shift budget.
The takeaway
Don't swap your account over on the strength of Meta's 12 percent figure. Run one clean, time-boxed test against your best manual campaign, measure new-customer share along with cost per purchase, and keep the manual structure that your learning and audience rules depend on. If the test wins, move budget in stages and keep a manual control running so you can see when performance changes.