Paid media
Why your ROAS dropped: the 7 real causes (and how to tell which is yours)
Your ROAS was at 3.5 and it has sat at 2.1 for three weeks. Nothing changed — or so it seems. Before touching bids, which is what everyone does first and is almost never the answer, you need to know which of these seven things is happening.
They are ordered by how often they actually occur, not by how interesting they sound.
1. Creative fatigue
By far the most common cause. Your ads burned out: people have seen them enough times to tune them out.
What it looks like: frequency climbing above 2.5 over a seven-day window, CTR falling steadily, CPM holding flat. That is the classic signature.
What to do: new creative, not cosmetic variations of the same. Different angles, different hooks, different formats. Changing the color of on-screen text resets nothing.
2. Seasonality and auction competition
Your CPM went up because more advertisers are bidding for the same attention. Black Friday, sale seasons, election cycles, major commercial dates.
What it looks like: CPM rises noticeably while CTR and conversion rate hold. Your ads are working just as well; you are simply paying more to show them.
What to do: often, hold. If your margin allows it and volume rises, a lower ROAS on more total sales can leave more absolute profit. This is a margin decision, not a ROAS decision.
3. Broken measurement
You updated the Shopify theme, installed a new app, the cookie banner changed, the conversions API broke. The ROAS you see dropped, but real revenue did not.
What it looks like: a widening gap between what the platform reports and what your store reports. This is always the first thing to rule out.
What to do: compare Shopify revenue against attributed revenue for the same period. If the gap opened exactly when ROAS “fell”, you have found the problem — and it is not in the campaigns.
4. Changes to the offer or price
The promotion ended, shipping went up, the first-purchase discount expired, the hero product that pulled everyone in went out of stock.
What it looks like: CTR holds but site conversion rate falls. People are still clicking; what stopped working is what they find when they land.
What to do: compare product page conversion rate against the same period last month. If it dropped, the problem is the offer, not the campaign.
5. Fragmented campaign structure
You kept adding ad sets and now twelve of them compete for the same audience, each with too few weekly conversions. None accumulates enough signal to exit learning.
What it looks like: many ad sets under 50 conversions per week, erratic day-to-day performance, high audience overlap.
What to do: consolidate. Fewer ad sets with more budget each. It is boring and it works.
6. Audience saturation
You have already sold to the easy part of your market. Who remains is harder to convince and more expensive to acquire.
What it looks like: prospecting ROAS falls while remarketing holds. The share of new customers in total orders declines month over month.
What to do: new angles for new audiences, not more budget behind the same message. This is where you find out whether your product has one use case or several.
7. Platform changes
Meta changes something — attribution model, conversion window, targeting policy — and the number moves without your business moving.
What it looks like: a sharp drop on a specific date, no prior signs of decay, and other advertisers reporting the same.
What to do: confirm against your store data before reacting. If real revenue is intact, what changed is the report.
The correct diagnostic order
Always the same, always in this order:
- Did real revenue drop? Check Shopify, not ads manager. If it did not, you have a measurement problem, not a campaign problem.
- Did CPM rise? If so, it is the market. Margin decision.
- Did CTR fall with frequency up? That is creative fatigue. Produce.
- Did site conversion fall? That is the offer or the store, not the campaign.
- None of the above? Now, and only now, look at structure and audience saturation.
The expensive mistake is starting at step five. Touching campaign structure when the problem was creative costs you two weeks of algorithm relearning and fixes nothing.
What not to do
Panic-cutting budget. It reduces conversion volume, degrades signal and makes the drop look worse. It becomes self-fulfilling.
Changing everything at once. Move creative, audiences, budget and structure on the same day and you will never know what fixed it or what made it worse.
Optimizing for ROAS above all else. A 6x on $2,000 of spend leaves less money than a 2.8x on $20,000. ROAS is an efficiency indicator, not a profit one.
If you have been going in circles on this diagnosis, QUÁNTIKA audits accounts spending over $2,500 USD a month. Book a call and we will go through it together.