Marketing10 min read
Advantage+ sales campaign ROAS dropped? Run these 7 checks first
admin
Monday, 10am. Your Advantage+ sales campaign ROAS dropped from last week's level, it carries most of your prospecting budget, and someone in the standup says "let's rebuild it as manual."
Hold that thought. When an Advantage+ sales campaign's ROAS drops, the cause sits in one of two places: the number is wrong, or the campaign changed. Rebuilding fixes neither if you have not found which one it is.
This guide gives you 7 checks, in the order we would run them, so you find the cause before you touch the campaign.
An Advantage+ sales campaign's ROAS usually drops for one of seven reasons: broken tracking, a shift in how much spend goes to existing customers, a learning reset, a fast budget increase, tired creatives, higher auction costs or ads selling stock you cannot ship. The first two can make ROAS look worse (or better) than it is. The rest change what the campaign delivers.
| # | Check | Where to look | What it tells you |
|---|---|---|---|
| 1 | Tracking | Meta purchases vs Shopify orders | Whether the drop is real |
| 2 | Existing customers | Audience segments breakdown | Whether old ROAS was inflated by repeat buyers |
| 3 | Learning reset | Campaign edits, migration date | Whether the campaign is relearning |
| 4 | Budget jump | Budget history | Whether you bought pricier reach |
| 5 | Creative fatigue | Frequency, CTR by ad | Whether the winners are tired |
| 6 | Auction costs | CPM trend | Whether everyone is paying more |
| 7 | Stock and delivery | Inventory, COD and RTO | Whether ads sell what you cannot deliver |
Run them in this order. The first two cost nothing to check and stop you from fixing a campaign that was never broken.
Start by confirming the drop exists in Shopify, not only in Ads Manager. If Shopify orders held steady while Meta purchases fell, the problem is measurement, not the campaign.
Compare the same dates in both tools:
If tracking checks out, the drop is real. Move on.
Look at the audience segments breakdown before anything else in the campaign. If last month's strong ROAS leaned on repeat buyers and that share has now fallen, the campaign may not have got worse at finding new customers at all.
Ads Manager can break a Sales campaign's results into new audience, engaged audience and existing customers (Jon Loomer). You define who counts as engaged and existing through custom audiences in your ad account settings, such as website visitors, customer lists or catalogue activity (Jon Loomer). If those audiences are stale or missing, the breakdown is wrong too.
Three facts make this check matter:
One attribution vendor's study of 55,661 campaigns found average new-customer acquisition cost on Advantage+ rose from $257 in May 2024 to $528 in May 2025, while manual campaigns improved over the same period (Wicked Reports, Jun 2025). Treat it as one vendor's data, not a rule. It does show why total ROAS can hide what you pay for a new customer.
For the full fix, see how to exclude existing customers from Meta ads.
Check the campaign's change history for the week the drop began. A migration from the old Advantage+ shopping campaign, or a significant edit, sends delivery back into learning, and results usually wobble while it relearns.
Meta's developer documentation is blunt about migration: moving a legacy campaign to the new Advantage+ setup "will force the campaign into the learning stage," and "there is not a workaround for the learnings" (Meta for Developers). The same page says Advantage+ sales campaigns can now hold more than one ad set, and that API v26.0 blocks all edits to legacy campaigns. Meta released Graph API v26.0 on 29 July 2026 (Meta changelog), so many accounts have migrated in recent months.
Meta's guidance, as relayed by practitioners, is that ad sets usually exit learning after about 50 results in the week after the last significant edit (Hussain Raniwala). If your drop lines up with a migration or a big edit, give it time before you edit again. Each new edit restarts the clock. Our guide to Meta ads learning limited covers what counts as a significant edit and the budget maths.
Compare the budget history with the ROAS line. If ROAS fell the week you pushed budget up sharply, you are probably paying more for each extra sale rather than looking at a broken campaign.
This is one of the most common questions merchants ask. One Shopify merchant scaled an Advantage+ campaign from $2,000 to $5,000 a day and asked why results got worse (Shopify Community, Mar 2025).
Meta's own bid strategy documentation explains part of it. With lowest cost bidding, Meta notes you have no control over cost, and costs can rise as the cheapest opportunities run out (Meta for Developers). More budget buys the next, pricier slice of the audience.
Say a brand spends ₹1L a day at a 3.0 ROAS and moves to ₹1.5L a day. If total revenue goes from ₹3L to ₹4L, blended ROAS reads 2.67. The extra ₹50,000 earned ₹1L, a marginal ROAS of 2.0. Those are illustrative numbers. The point is to judge the extra spend on what the extra spend returned. How to scale Meta ads without losing ROAS covers how far to raise budgets.
Sort ads in the campaign by spend and look at frequency and CTR over the last 2 to 3 weeks. If the 2 or 3 ads taking most of the spend show rising frequency and falling CTR, creative fatigue is the likely cause.
Advantage+ tends to pile spend onto early winners. One Shopify Community thread describes CTR dropping and CPM rising after a week or two in an Advantage+ campaign with 50+ product creatives, and a reply describes spend concentrating on the first winners (Shopify Community, Dec 2025).
Fix it by adding new angles, not small tweaks to the same ad. Our guide to creative fatigue vs audience saturation shows how to tell the two apart.
Check CPM next to ROAS. If CPM rose and CTR and conversion rate held steady, the campaign is fine and the auction got more expensive.
This happens across the market, not only in your account. Meta reported that ad impressions grew 10% and the average price per ad grew 14% year on year in the third quarter of 2026 (Meta investor relations, Oct 2026).
A market-wide CPM rise calls for a new ROAS target or better margins, not a campaign rebuild. See why Meta CPMs rise in India to separate market cost from account problems.
Check stock and delivery last, because it hides well. Advantage+ will keep spending on a product that sold out, and COD orders that never get delivered still count as purchases in Meta.
Three things to check:
Not as a first move. Switching resets learning, and if the cause was tracking, existing-customer mix, fatigue or market CPMs, a manual campaign inherits the same problem.
If the 7 checks find nothing, test instead of switching. Run a manual campaign next to Advantage+ with a fixed share of budget for 2 to 3 weeks and compare them on cost per new customer, not total ROAS. Keep the winner. The Wicked Reports study above is a reason to measure new-customer cost carefully, not proof that manual always wins.
Most of these checks need data from more than one place: Meta, Shopify, inventory and courier reports. That is why a drop often takes days to explain.
This is the kind of leak Clevrr AI, an AI performance marketer for D2C brands, watches for every morning across ads, store, stock and deliveries. It splits each revenue change into four drivers (demand, order value, operational loss and wasted ad spend) and prices each in rupees. BetterHood recovered about ₹2.7L in a month that was going to retargeting existing buyers. ShaperCult had spent ₹1.6L over four days on ads for out-of-stock sizes; spend now pauses automatically on stock-outs. Clevrr respects learning phases before it scales or pauses anything.
When an Advantage+ sales campaign's ROAS drops, check the number before the campaign. Confirm tracking, then look at the existing-customer share, then learning, budget, creative, CPM and stock, in that order. Most drops have a cause you can find in an hour, and most do not need a rebuild.
If you want these checks run on your account every morning, start free with 500 credits.
No. A few days of lower ROAS is often noise, a learning reset or higher auction costs, and turning the campaign off throws away what it has learnt. Check tracking and the audience segments breakdown first. Act only when the drop holds for a week or more and you have found a cause you can fix.
Wait until the campaign exits learning before judging it. Meta's guidance, as practitioners relay it, is that ad sets usually exit learning after about 50 results in the week after the last significant edit. Avoid stacking edits during that time, because each significant edit restarts learning and makes the results harder to read.
Probably some of it. Open the audience segments breakdown in Ads Manager to see spend and ROAS for new, engaged and existing customers. Make sure your existing-customer audiences in ad account settings are current. If too much goes to repeat buyers, Meta suggests a separate ad set for existing customers with a spend limit.
Not by itself. Minimum ROAS bidding tells Meta to look for purchases above your target, but Meta warns that a target set too high may lead to under-delivery. It can protect efficiency while costs rise, at the price of volume. Find the cause of the drop first, then decide whether a ROAS floor fits.
Join hundreds of D2C brands using Clevrr AI to automate their growth and efficiency.
Join 500+ D2C founders using Clevrr AI to automate their operations and scale faster.
Get Started Free
