Why your ROAS is dropping, and how to improve ROAS without spending more

Here is a call we know well. ROAS was 3x last month. This week it is 1.8x. The owner has already paused two ad sets, cut the budget, and duplicated the "winner." Now four things are broken instead of one.

So before any fixes, one piece of math. ROAS is revenue divided by ad spend, and it breaks into three parts:

Is my ROAS actually dropping, or is it the reporting?

Check this first, because it is the cheapest to rule out and it fools the most people.

Conversion delay. Google's own help docs say recent performance can look weaker than past performance because some people who clicked haven't bought yet, while the spend is already fully reported. Conversions can be credited up to 90 days after the click, depending on your conversion window. So comparing the last three days to last month is rigged against the last three days. Compare two complete periods instead.

Broken or changed tracking. A theme update that dropped the pixel. A new checkout app. A duplicate purchase event that got "cleaned up." A changed attribution setting. Put platform conversions next to the orders in Shopify or your CRM for the same days. If real orders held steady and platform conversions fell, you don't have a performance problem. You have a tracking problem, and every edit you make to the ads is a guess.

Did you reset the learning phase?

Probably, if you made changes in the last week. This is the most common self inflicted ROAS drop we see.

Meta says an ad set exits learning once performance stabilizes, which typically happens around 50 optimization events after the last significant edit, and that cost per action can be worse while it is learning. Meta counts pausing an ad set, or changing the optimization event, audience, or creative, as significant. Budget and bid changes can count too, depending on how big they are.

Google works the same way. Its current guidance says Smart Bidding can take up to roughly 50 conversion events or three conversion cycles to calibrate after a change. On a home service account doing 20 leads a week, that is weeks, not days.

Our take: one bad Monday is not a reason to edit five ad sets. Make one change, then leave it alone long enough to read it.

Is your creative worn out?

Look at the Delivery column in Meta Ads Manager. Meta flags Creative fatigue when an ad's cost per result is at least twice what it used to be, and Creative limited when it is higher but not yet double. If you see either, that is your answer.

The fix is new ads that are actually different (new angle, new format, new person on camera), added to the ad set that already works. Not a rebuilt campaign. Small audiences and one ad per ad set wear out fastest. And don't diagnose fatigue from frequency alone. A high frequency with steady results is fine.

Did the auction just get more expensive?

If cost per click or CPM jumped while your click through rate and conversion rate held, it isn't you. It's the auction. Q4 is the classic case: from Black Friday through the holidays, every retailer piles into the same inventory and prices rise. On Google, open Auction insights and look for a new name with a rising impression share.

You can't out bid a seasonal auction profitably. You can protect margin: tighten to your best products and best audiences, and raise order value so each click is worth more.

Did the site or the offer stop converting?

If clicks cost the same and people still click, but fewer of them buy, the problem is after the click. The usual suspects: the best seller went out of stock, a price or shipping threshold changed, a promo ended, or a theme update slowed the product page. Our Shopify conversion checklist covers the 19 fixes we check on that side.

How to improve ROAS without spending more

Here is the order we work a ROAS drop on a new account:

1. Reconcile the numbers. Platform conversions against real orders or booked jobs, over two complete periods. Rule out reporting first.
2. Run the formula. Conversion rate, order value, cost per click. Find the one that moved.
3. Freeze edits for a week. Let anything in learning finish learning.
4. Add creative, don't rebuild. Three to five genuinely different ads into the ad set that was working.
5. Raise order value. Bundles, a free shipping threshold just above your average order, a post purchase offer. ROAS goes up without touching the ads.
6. Scale only behind efficiency. Budget follows results, never leads them.

That last point is how the NUDE Envie account went. We took it over near break even at 1.07x ROAS in February 2026. By July it was at 1.87x and cost per order had dropped from $82 to $32, and the budget on winning campaigns only went up (48%, from $300 to $445 a day) as efficiency climbed. If you want us to look at your account the same way, start with our Meta Ads or Google Ads management pages, or see how we run e-commerce ads as one system.

One more honest thing. A falling ROAS is not always bad news. Your break even ROAS is 1 divided by your gross margin, so a product with a 50% margin breaks even at 2.0x. Dropping from 4x to 3x while you double spend can mean you are making more money, not less. Know your break even number before you call it a crisis.

FAQ