Advantage+ sales campaigns: what changed and how to test vs manual

The question lands every week: "Should we turn on Advantage+ or keep it manual?" First, a naming fix. Advantage+ Shopping no longer goes by that name. Meta now calls it an Advantage+ sales campaign, and the same approach applies to leads and app campaigns. In practice, Meta takes over three levers (audience, placements and budget) and you move from giving orders to giving suggestions. You can absolutely compare it against a manual campaign, but only if the test is fair: same event, same attribution setting, same creatives, and revenue checked outside the platform.

What changed from Advantage+ Shopping to Advantage+ sales

Three changes matter if you run accounts:

1. The name and the scope. On Meta's Advantage+ page, sales campaigns are described as the former Advantage+ Shopping campaigns, listed next to Advantage+ app and leads campaigns. It is no longer a separate, ecommerce-only campaign type.
2. The structure. Instead of a special campaign type, Advantage+ is now a state the campaign is in: it counts as Advantage+ when Advantage+ audience, Advantage+ placements and Advantage+ campaign budget are on.
3. The API. Meta's developer documentation says that as of Marketing API v25.0, developers can no longer create Advantage+ Shopping campaigns the old way. API-built campaigns now use Advantage+ audience, budget and placements instead.

For leads, Meta is explicit: every campaign with the leads objective in Ads Manager has Advantage+ on by default.

When a campaign stays Advantage+

Meta spells out the conditions on its leads campaign page, and they show the three-lever logic clearly:

• Budget: Advantage+ campaign budget stays on.
• Audience: Advantage+ audience is on in at least one ad set. That means the default audience, or inputs used only as suggestions. Age starts at 18, any location or language, and only custom audience exclusions.
• Placements: Advantage+ placements on in every ad set, with all placements selected.

Account-level audience and placement controls do not switch Advantage+ off. A hard control at the ad set level can. Meta's own advice is to use controls only for genuine business restrictions.

Detailed targeting is now a suggestion

This is the part that trips people up. Interests, age, gender, custom audiences and lookalikes did not disappear. They became audience suggestions. The system prioritizes people who match the suggestion and then expands beyond it when it finds results.

What that means day to day:

If your business has a real restriction (you only ship to three states, the product is 21+), use a control. If it is a preference, use a suggestion.

What you still control

Quite a lot, and it is what moves results most:

• The conversion event. Optimizing for purchases, leads or qualified leads changes everything. A bad event teaches the system to find the wrong person faster.
• The creative. With broad audiences, the ad decides who it speaks to. Meta Ads automation vs an AI assistant breaks down what stays with Meta and what stays with you.
• The offer and the landing page. No automation fixes the wrong price or a slow page.
• Location, language and exclusions. Including excluding current customers on acquisition campaigns. In October 2026 Meta announced that Customer Lifecycle Strategy is available to all advertisers: when a campaign is marked as acquisition, it recommends how to exclude existing customers.
• Total budget. How much goes in is your call. How it splits across ad sets is Meta's.

Meta publishes cost-per-acquisition improvement figures for Advantage+ sales campaigns on its own product page. Those are Meta's own numbers, with no method shown on the page. Treat them as a hypothesis to test in your account, not a promise.

How to test Advantage+ vs manual without fooling yourself

The classic mistake is comparing the ROAS of an Advantage+ campaign that is selling to people who were going to buy anyway with the ROAS of a manual cold-audience campaign. Advantage+ "wins," and nobody notices that part of those sales would have happened regardless.

A fair comparison follows five rules:

1. Same optimization event and same attribution setting. If one uses 7-day click and the other 1-day click, the test is over before it starts. Attribution windows explained covers the difference.
2. Same creatives. Otherwise you are testing creative, not structure.
3. Same dates and similar budgets. Running at the same time, not one in September and one in October.
4. Use Ads Manager's A/B test when you can. It splits the audience so the two campaigns do not compete for the same people. The setup is in Meta Ads A/B testing.
5. Separate new customers from existing ones. Meta offers a results breakdown for new audience, engaged audience and existing customers once you define those audiences (from website, app, customer list or catalog data). Without that split, the campaign that sells most to people who already buy from you looks like the best one.

Above all, check revenue outside the platform. Your CRM or payment processor shows whether the order came in, whether it was a new customer and whether it was refunded. When Meta's number and your bank account tell different stories, the bank account wins. To measure the true effect of a campaign, the most solid route is an incrementality test.

When to stay manual

Advantage+ is not mandatory for everything. A manual setup still makes sense when:

• Conversion volume is too low for the system to learn.
• You have a real audience restriction that a suggestion will not respect.
• You are testing one specific thing (an audience, an offer) and need to isolate it.

Outside those cases, the useful question is not "Advantage+ or manual" but "what do I keep as a control and what do I leave as a suggestion."

Your next step

Open your sales campaign and check, ad set by ad set, what is set as a control and what is a suggestion. Then run an A/B test with the same event and the same creatives, and judge it by your CRM. To compare Meta, Google and TikTok in one place, see the plans: from US$ 31.20 a month on the annual plan, with the first 7 days for US$ 2.99.