Should you let Meta Advantage+ run your campaigns? The honest answer has two parts: it depends on your account, and the number Meta uses to convince you is softer than it looks. Meta reports a 12% lower cost per action and a 15% higher return on ad spend on average, according to Social Media Today (2022) reporting Meta’s own figures, but that claim covers e-commerce advertisers and carries no published methodology. Independent measurement points the other way: across 640 incrementality experiments, manual campaigns won in 58% of cases, according to Haus (2025), a US incrementality vendor whose panel skews to e-commerce.
This article does not tell you to switch everything on or run away. It does the thing most guides skip. First it clears up the confusion at the heart of the topic, the difference between the Advantage+ campaign and the Advantage+ features. Then it reads Meta’s claim, the real conversion thresholds, and the independent data with the same scepticism, and asks what changes for a B2B advertiser with a long cycle and few conversions.
Advantage+ Sales or Advantage+ features: clear this up first
The single biggest source of confusion is that “Advantage+” names two different things, and people argue past each other because of it.
Advantage+ Sales is a campaign type. It is the end-to-end automated campaign, formerly called Advantage+ Shopping, that hands targeting, budget allocation and placement to Meta’s algorithm. Advantage+ features are separate toggles you can switch on inside a normal manual campaign: Advantage+ Audience, Advantage+ Placements, Advantage+ Creative and Advantage+ campaign budget. You can use the features without ever running the automated campaign.
So “should I use Advantage+” is really two questions. One is whether you hand a whole campaign to automation. The other is which individual decisions you delegate inside a campaign you still control. Those questions have different answers, and conflating them is the first mistake.
One practical note on naming. Meta renamed Advantage+ Shopping to Advantage+ Sales in early 2025, widened it beyond e-commerce to sales, leads and app installs, and brought back ad sets and some audience controls, according to PPC Land (2025). Same automated engine, broader remit, a few more controls returned to your hands.
What does Meta claim, and how solid is it?
Meta’s headline number is real, and it is also weaker than it sounds. The company reports that advertisers on Advantage+ saw a 12% lower cost per action and a 15% higher return on ad spend on average, a figure that first appeared at the October 2022 launch and is still repeated across the ecosystem, per Social Media Today (2022).
Three caveats travel with that claim, and Meta does not print them next to it. The measurement covers e-commerce advertisers, not B2B services. It comes with no published sample size or protocol, so you cannot judge how it was run. And the comparison baseline is Meta’s own manual campaigns, which the platform has every incentive to frame favourably. A number without its method is not proof, it is a well-dressed assertion. It is a signal worth noting, not a verdict, and it does not describe your account any more than an average Facebook ads cost describes your CPM.
How many conversions do you need for Advantage+ to work?
This is the question that quietly decides the whole debate, and it has a concrete answer. Automation needs signal, and signal means conversions. Plan for roughly 50 conversions per ad set per week, the volume Meta’s learning phase needs to stabilise delivery, a threshold repeated across practitioner sources such as Wonderful (2026). Below it, the algorithm keeps relearning, and your cost per result stays volatile.
That threshold turns into a budget floor. Our calculation: at a target cost per result of 50 US dollars, reaching 50 weekly conversions needs about 2,500 US dollars per ad set per week (50 times 50). At a 200 US dollars cost per qualified B2B lead, the same maths asks for 10,000 US dollars per week just to feed one ad set. This is an illustration of the rule, not a figure Meta publishes, and it is the reason the automated campaign fits high-volume accounts and starves low-volume ones.
The 50-per-week number is a guideline, not a hard gate. Some accounts run below it and still perform, they simply carry more variance. But as a planning input it is the honest test of whether the fully automated Advantage+ campaign can even learn on your account, before you argue about whether it should.
Is Advantage+ better than manual campaigns?
On average, and by the strictest available measurement, no. This is where independent data flatly contradicts Meta, and the contradiction is the point. Across 640 geo-based incrementality experiments over 18 months, on brands spending on average about 1 million US dollars per month, Haus (2025) found that manual campaigns delivered higher incremental return in 58% of tests, against 42% for Advantage+.
The timing detail matters more than the headline. Advantage+ ran 9% ahead of manual at the experiment midpoint, then finished 12% behind, as summarised by AdExchanger (2025). Automation starts fast, manual wins the long game. By the end, manual returned about 12 US dollars more revenue per 100 US dollars spent, and Advantage+ over-reported its own contribution by roughly 12 percentage points versus what independent measurement credited it with.
Two guardrails on this data, because one bias does not cancel another. Haus sells incrementality measurement, so it profits when platform attribution looks wrong. And its panel is mostly e-commerce and DTC, not B2B. This is exactly the gap that opens up when GA4 and Meta’s numbers do not match: platform-reported results are not the same as incremental revenue, and only your own measurement settles it.
When Advantage+ complements manual, and when it cannibalises
There is no universal winner, only contexts. The data and practitioner consensus sketch two profiles. Read the table as a starting point for your own test, not a ruling.
| Context | Tends to favour | Why |
|---|---|---|
| High weekly conversion volume | Advantage+ Sales | The algorithm gets enough signal to learn fast |
| Clean, abundant first-party data | Advantage+ Sales | Good signal feeds the automation properly |
| New account, low volume | Manual | Too little signal for the algorithm to converge |
| Narrow B2B or niche audience | Manual | Broad targeting dilutes into off-target profiles |
| Small budget | Manual | Little room for the exploration phase |
| Strict lead qualification (B2B) | Measure first | Cheaper leads, weaker proof they are qualified |
The cannibalisation risk is specific: run an automated Advantage+ campaign next to tightly controlled manual campaigns, and the automated one can absorb your cheapest existing conversions, then report them as its own incremental wins. That is how a dashboard shows a lift that your revenue never sees, and it is a cousin of the problem you meet when ROAS drops as you scale. The features are lower risk. Jon Loomer’s 30-day test found Advantage+ Audience matched or beat detailed targeting and clearly beat lookalikes at a lower CPM, per Jon Loomer Digital (2025), so delegating the audience lever is often a safer first step than delegating the whole campaign.
What should B2B expect versus e-commerce?
A different starting point, and a different verdict. Every headline number in this debate, Meta’s 12% and Haus’s 58% alike, comes from e-commerce and DTC panels. A B2B advertiser sits on the opposite side of two structural facts.
The first is volume. A long sales cycle produces few weekly conversions, so the automated campaign rarely clears the 50-per-week learning threshold, and it never learns properly. The second is fit. In B2B, more volume is not the goal. High-volume, broad-targeted delivery tends to produce lower-intent, lower-fit leads, which a long cycle then makes expensive to chase, as Search Engine Land (2024) argues. Cheaper leads that never qualify cost sales time, not just media budget.
The practical read for B2B: treat the Advantage+ features as a real option and the fully automated Sales campaign as a claim to verify. Advantage+ Audience and Placements can help a narrow account explore without losing the plot. The automated campaign, on a low-volume account optimising for qualified pipeline rather than raw leads, is exactly the case where you measure before you trust. It is the kind of call we work through in B2B paid acquisition: establish what is incremental before handing anything to the algorithm.
In short
- Separate the campaign from the features. Advantage+ Sales is a whole automated campaign; Advantage+ Audience, Placements, Creative and budget are toggles you can add to a manual campaign one at a time. They are two decisions, not one.
- Do not decide on an average. Meta’s 12% and 15% come from an e-commerce test with no published method. Haus’s 58% comes from an incrementality vendor with an e-commerce panel. Neither describes your account, and Advantage+ tends to over-report what it actually adds.
- Match automation to your signal. The automated campaign needs about 50 conversions per ad set per week to learn. High-volume e-commerce can feed it; a long-cycle, low-volume B2B account usually cannot, and should lean on the features while measuring the rest.
Weighing whether to hand your Meta campaigns to the algorithm, and want the call made on your own numbers rather than an industry average? Book a diagnostic: we establish what is genuinely incremental on your account before choosing between automation and manual control.