How to Improve ROAS: Why Your ROAS on Meta Ads Is Dropping and What You Can Do Now

Daniel Huber, MBAFounder of Plaith

Updated · 10 min read

Image created with AI

I've been reading about what problems are out there again. And this morning two new posts were right at the top, both describing the same feeling: the number everyone is watching is going down. ROAS.

Anyone who googles improve ROAS usually wants one thing: for it to be like last week again. The three cases below show that very different causes can sit behind a falling ROAS. And that the fastest reaction is often the most expensive one.

In Short: What ROAS Measures and What It Doesn't

ROAS stands for return on ad spend. You divide the revenue Meta attributes to your ads by what you spent on the ads. A ROAS of 2 means: for every euro of advertising, Meta reports two euros of revenue.

ROAS does not measure two things. First, your profit: cost of goods, shipping, fees and returns come after that. Second, the whole truth: Meta only counts what it can attribute, and only within an attribution window. As things stand now, the usual setting is "7 days after a click or 1 day after viewing the ad". In 2026 Meta removed the longer view windows from reporting, so the numbers in your account may have shifted compared with before. You can see which setting applies in your account in Ads Manager. That's why ROAS is an important signal, but not a verdict.

Case 1: ROAS Collapses Every Evening

On X, an e-commerce seller asks around whether anyone else is seeing the same thing (to the original post). By his account, on the third day in a row his ROAS drops from over 2.2 to under 1.8 in the last six hours of the day. Morning and midday are fine, he says, and in the evening it goes downhill.

That looks like a clear pattern, and many people would switch the ads off in the evening right away. Before you do, check two things:

  • Is it real or just not counted yet? Meta assigns a purchase to the day of the ad, not the day of the purchase, at least in the usual reporting view. Someone who clicks in the evening and only buys the next morning shows up for the previous day later. If you look at the same evening, the evening can look worse than it ends up being. So compare completed days, not the current one.
  • Is it a real time-of-day pattern? Look at several weeks broken down by hour and compare with the orders in your shop, not just with what Meta reports. Only when the pattern is stable in both sources over a longer period is it worth reacting to.

The lesson: Three days are a hint, not proof. And ad scheduling at Meta was tied to a lifetime budget for a long time. Check in your Ads Manager which options you are offered, and expect that it may be a rebuild, not a small adjustment.

Case 2: More Revenue, Less Profit

The second post comes from a shop founder who is publicly documenting a challenge: the goal is $10,000 in revenue per day (to the original post). On day 72 he reports a new revenue record, almost $8,000. Orders are rising. But profit has dropped slightly, he says, and ROAS has fallen again. His own conclusion: he wants to be more careful about how much he spends on scaling.

That isn't necessarily an error in the account, but often a familiar pattern. With every additional euro, Meta often reaches people who are a bit less ready to buy than the first ones. The average ROAS then falls, even when everything is done right. So the decisive question is not "How do I hold my ROAS?" but "Does the next euro still bring profit?".

The lesson: If you want to scale Facebook Ads, you need a floor below which growth no longer brings profit. How to scale without an exploding CPA is covered in detail in Scaling Facebook Ads.

Case 3: From 3.2 to 0.8 in Four Days

The third case is the most dramatic (to the original post). A men's T-shirt brand from India advertises, by its own account, with around $50 a day, and the product is running at 40 percent off. The week before, ROAS was 3.2. In the last four days it was 0.8, with only two purchases. CPM and click price have risen only slightly, but add-to-cart actions are going down. Nothing was changed in the shop. The founder asks: wait, or rebuild now?

The best answers in the thread are remarkably calm. If all campaigns collapse on the same day, then in the view of those replying it is rarely down to the structure. One user suspects customers are waiting for bigger discounts before the Diwali festival. His advice: don't rebuild the account, let the best campaign keep running with less budget, add a new ad set with fresh creatives. When the founder duplicates the old winning ad set, the next hint comes: two ad sets with the same ads compete with each other, and with a budget this small two ad sets are enough. And then don't touch anything for a few days, because with two purchases a day, a single day tells you nothing.

These are assessments from a forum, not established facts. But they fit what many experienced advertisers say.

The lesson: A drop that hits all campaigns at the same time points more to the market, the season or tracking than to your ads. A rebuild also resets the learning phase and throws away exactly the learning you need when things pick up again. You'll find more on sudden drops in Facebook Ads Don't Work Anymore.

Why ROAS Drops: The Five Most Common Causes

From these and many similar cases, five causes emerge that you should check in order:

  1. Tracking and attribution. If purchase events are missing, ROAS is artificially small. Compare Meta's numbers with the real orders. How to set this up cleanly is in Setting Up the Meta Pixel.
  2. Market and season. Holidays, discount weeks and competitors with more budget shift prices and willingness to buy. If everyone is affected, you are too.
  3. Tired creatives. When the same ads run for weeks, their effect often drops. An early warning sign is falling add-to-cart actions at constant click prices, as in Case 3.
  4. Scaling. More budget often reaches people who are less ready to buy. ROAS drops even when everything works.
  5. The offer. If the discount no longer stands out because everyone is offering one, even better ads help only to a limited extent. And if the purchase decision stalls on the page, take a look at Improving Landing Page Conversion Rate.

Improving ROAS in Five Steps

Here is how to proceed when ROAS drops, without turning the whole account upside down:

1. Know Your Break-Even ROAS

Before you get annoyed about a ROAS of 1.8, you need to know whether 1.8 means a loss or a profit for you. Calculating break-even ROAS is simple: divide 1 by your margin after cost of goods, shipping, fees and returns. If you keep 50 cents of every euro of revenue, your break-even ROAS is 2. If you keep 25 cents, it's 4. Only with this number do you know whether you really have a problem. If you calculate with the margin before shipping and fees, you set the break-even too low. And break-even is not your goal: if you want to make a profit and also cover fixed costs, your target ROAS has to be above it.

2. Zoom Out Before You Intervene

Evaluate completed days and at least a week, not individual hours. Especially with a small budget and few purchases, the numbers swing a lot.

3. Compare Meta With Your Shop

How many orders do you really have, and how much have you spent on advertising in total? If the shop is stable and only Meta reports less, the problem lies more in the measurement than in the ads.

4. Change One Thing, Not Five

New creatives in an existing ad set, an adjusted offer or a budget correction. But not everything at once, otherwise you won't know afterwards what helped. How to test creatives fairly is covered in Creative Testing on Meta Ads.

5. Give Every Change Time

After a change, delivery needs a few days to stabilize. Meta distinguishes between small and significant changes: significant changes, such as to the audience, creative, optimization event or larger budget jumps, can restart the learning phase. If you rebuild every day, you reset the learning again and again. Which changes exactly count, and from which budget jump, is explained in Meta's help on the learning phase.

The Most Common Mistakes When ROAS Drops

  • Rebuilding the account even though the whole market is weak.
  • Duplicating winners until several ad sets with the same ads address the same audience and split delivery and learning.
  • Deciding by the hour even though the day's purchases are not yet fully counted.
  • Looking only at ROAS instead of profit after all costs.
  • Doubling the budget in a record-high rush without knowing from which point the next euro brings a loss.

How much budget makes sense at all, you can work out with our guide Calculating Your Ad Budget.

What You Get Out of It

Maybe you know this: the first thing you do in the morning is open Ads Manager, before the coffee is even ready. The number is red. You change something, look again at noon, change something else. And by evening you no longer know whether you just fixed something or broke it.

If you know your break-even ROAS and check in a fixed order, panic turns into a decision.

Here is what you can gain from it:

  • Calm when the number fluctuates. You know from which value you really have to act, and from which you don't.
  • Less money burned on panic changes. You don't throw learning away just because one evening looked bad.
  • Growth that is more likely to pay off. You scale up to the point where the next euro still brings profit, and not beyond.

Nobody can seriously promise you a guaranteed ROAS. It depends on your product, your margin and your market. But you make your decisions based on numbers, not on fear.

What This Looks Like at Plaith

Plaith helps self-employed people and small businesses win customers through Meta Ads. At Plaith, 6 specialists, combining people and AI, take care of videos, ad messaging and campaign management, from €26 a day plus a one-time setup. That means: regularly new creatives in testing and campaign management that doesn't fall into a frenzy after every bad evening.

If you want to know where your ROAS is currently stuck, book a call with René. We look at your numbers together, and if it isn't a fit, we'll tell you honestly.

Frequently Asked Questions

What Is a Good ROAS for Facebook Ads?

That depends on your margin. A good ROAS is above your break-even ROAS, meaning the value from which you make a profit after all costs. Work it out by dividing 1 by your margin.

Why Does My ROAS Drop When I Increase the Budget?

Because with more budget Meta often also reaches people who are less ready to buy. The average ROAS then often drops, even if the campaign is running well. What matters is whether the additional revenue still brings profit.

Should I Pause Ads When ROAS Collapses?

Usually not right away. First check whether all campaigns are affected, whether the tracking is right and whether you are comparing completed days. It is often better to let the best campaign keep running with less budget and to test new creatives in a targeted way.

Can I Run Facebook Ads Only at Certain Times of Day?

Yes, that works through scheduling in the ad set. It was tied to a lifetime budget for a long time, so check in your Ads Manager which options you are offered. And first check over several weeks whether the time-of-day pattern is really stable in Meta and in your shop.

Read next