Journal · Playbook
Why your Meta ad ROAS dropped and what to check before you cut anything
Meta changed how it counts click-through conversions in March 2026. Reported ROAS is down 15 to 30 percent for most campaigns. The ads probably did not get worse. The measurement got more honest.
A used-car lot in Westfield has been running Facebook lead ads for fourteen months. Cost per inquiry was $18. Decent. Last month it jumped to $31 and the reported ROAS dropped from 4.8 to 2.9. Same budget. Same creative. The owner cut the campaign before I could call him back.
He did not need to.
Meta changed how it counts conversions in March 2026. If your reported ROAS (return on ad spend: the revenue you got back for every dollar you put in) has dropped over the last few months without any obvious reason, this is almost certainly why. Here is what changed and what to check before you make any budget moves.
What Meta changed
For years, Meta counted any ad interaction as a "click" for attribution purposes. Someone liked your ad, then bought something within seven days: Meta logged that as a click-through conversion. The same went for shares, saves, comments, and video plays. Any touch point opened the seven-day window.
Starting in March 2026, only link clicks count. A user has to actually follow your ad to a website, landing page, lead form, or other destination. If they liked your ad and then converted, that now falls under a different category: engage-through attribution.
Your ads did not get worse. The measurement got more honest.
The math on the reported drop
If 25% of your attributed conversions previously came from engagements rather than actual link clicks, your reported ROAS just dropped by roughly 25%. Not because fewer people bought. Because those conversions are now counted in a separate column.
Meta estimates 15 to 30 percent drops in reported ROAS for most campaigns after this change. Remarketing campaigns tend to see the largest drops because they have more view-through and engage-through activity baked in. Cold-audience lead-gen campaigns tend to see smaller drops because more of their conversions were real link clicks to begin with.
That is a significant number on paper. Read the paper carefully before you act on it.
Two things to check right now
Your actual revenue. Open your Square, your CRM, your booking software, your call log. Is the real number of leads, appointments, or sales flat compared to two or three months ago? If yes, your campaign is performing the same. The reporting number is a reclassification, not a performance signal.
Your Ads Manager attribution breakdown. In your campaign results, you can now see conversions split into click-through, view-through, and engage-through. If your "missing" conversions reappeared in the engage-through column, that confirms it. Nothing disappeared. It moved.
If your actual revenue is flat and your engage-through column is populated, the campaign did not change. The report did.
One case that actually warrants a real look
Remarketing campaigns.
If you have been running ads to people who visited your site, interacted with your Instagram, or watched your videos, those campaigns were probably counting a lot of engage-through conversions as click-through under the old system. The ROAS numbers were likely inflated for years.
Now you can see a cleaner number. For some shops, that cleaner number is genuinely lower. Not a reporting artifact. A real signal that the remarketing audience was working less hard than the dashboard suggested.
If remarketing makes up more than 40% of your Facebook spend, pull the engage-through breakdown and look at it honestly. The audience or the creative may need adjustment.
For lead-gen campaigns targeting cold audiences, the picture is simpler. If people are clicking the ad and filling out the form, the new reporting is telling you the same thing the old reporting was telling you. You just have a cleaner number now.
The actual action
Do check your real revenue numbers before touching a campaign.
Do open your attribution breakdown and find where the "missing" conversions went.
Do look harder at any remarketing campaign that was delivering most of its ROAS through engage-through rather than link clicks.
Do not cut a campaign because a paper number dropped 20%.
Do not tell your Meta rep the platform stopped working. It did not.
Do not change creative or targeting in response to a reporting change.
The used-car lot in Westfield is back running the same campaign. Cost per actual inquiry: $19 once you strip out the engage-through noise. Basically the number he started with.
He lost six weeks of leads in the meantime.
Sources
| # | Source | What it backed up |
|---|---|---|
| 1 | Meta for Business: Simplifying Ad Measurement for a Social-First World | Meta's announcement that only link clicks count for click-through attribution starting March 2026 |
| 2 | Why Did My Meta Ads Performance Drop? — Three Chapter Media | 15 to 30 percent ROAS drop range for most campaigns; reclassification vs. performance decline |
| 3 | Meta Ad Attribution Changes: Link Clicks, Engage-Through and Video Rules Explained — ALM Corp | Breakdown of engage-through attribution column and how conversions are reclassified |
| 4 | Meta Ads Click Attribution: How Strategy Is Changing in 2026 — Five Nine Strategy | Impact on remarketing campaigns and ROAS inflation from engage-through attribution |
| All sources verified May 2026. |