When your return drops, the cause is almost never where you first look. It is usually one of three things: the account is chasing the wrong kind of conversion, the traffic is landing somewhere that does not sell, or you are comparing two platforms that count differently.
These questions sort out which one it is.
Start with what your account is chasing
This is the most expensive mistake in advertising and it never shows up in a performance report.
Someone, at some point, told the account that a newsletter signup counts as a conversion. Ever since, the bidding has been happily buying newsletter signups and reporting a wonderful cost per conversion, while nobody buys anything.
"What is my account actually counting as a conversion?"
You will often find three or four things counted, only one of which is a sale. Some campaigns may be set differently from the rest, so ask which ones.
The fix is to tell the account to count only what you mean by a sale, and which campaigns should switch to it. You can also say that a customer from one country or on one device is worth more than another.
Then look at where the traffic lands
"Which landing pages am I sending budget to, and which ones convert?"
This puts the ad numbers and what people did on the page side by side, which is the only way to end the argument about whether it is the ad or the website.
The follow-up matters more than the first answer:
"Is this a broken page, the wrong people, or the wrong promise in the ad?"
Three answers, three different people who need to fix something.
Comparing platforms without fooling yourself
"Compare Meta and Google for last month and tell me which deserves more money."
Putting two platforms side by side is misleading before you start, because they each decide differently who gets credit for a sale, and they are often doing different jobs. The comparison names those differences first rather than pretending the numbers are equivalent.
Your shop is the tie-breaker:
"Compare what the platforms say they sold with what the shop recorded."
These never match exactly. Knowing why is more useful than averaging them.
Then act, and check whether it helped
Once you know where the return is going, the fixes are ordinary: move money to what works, block the searches that never sell, cut the placements that waste, change where the ads point.
A month later:
"Did that change last month actually do anything? What would have happened if I'd left it alone?"
That is a different question from "did the number go up", which could just as easily be the season, a competitor pausing, or luck.
The order to do it in
"What is my account counting as a conversion?"
"Which landing pages get my budget, and which convert?"
"Which campaigns run out of budget every day but do well?"
"Move 20% of the budget to those. Show me before and after."
A month on: "Did that work?"
What it will not do
It does not know your profit margin. It works from the revenue the platforms and your shop report, not from what you actually make.
It does not report today.
It does not change how Google Analytics assigns credit for a sale. It reads that setting so you can see it.
It does not test whether your sale tracking fires. Use the platform's own tools for that.
When something looks wrong
What you see | Why | What to do |
Cost per conversion looks great and nothing is selling | Something cheap is being counted as a conversion | Ask what is being counted, then narrow it to real sales |
The platform and your shop disagree on revenue | Different credit windows, currency, or a tracking gap | Ask for the comparison and the reason, do not split the difference |
One channel looks far worse than the other | They may be counting differently or doing different jobs | Ask for the comparison that separates those out |
Return dropped and nothing changed in the campaigns | Usually the landing page or the tracking, not the ads | Ask whether it is the page, the audience or the ad |
Common questions
Does it work out ROAS on profit?
No. It uses the revenue your platforms and shop report. It does not know your margins.
Can it change what my campaigns optimise towards?
Yes on Google Ads, and you see the change before it happens.
Why do Google Analytics and Google Ads disagree?
Usually how credit is assigned, a reporting delay, or a conversion that is not being passed across. All three are checkable.
Can it prove a change worked?
It measures against what would likely have happened otherwise. That is a measurement, not a promise.
Still need help?
Message us in the app or write to support@orphex.co. Never share your password or an access key with anyone, including us.