The conversion tracking mistake in most ad accounts
If your conversion action is a page view, the algorithm is optimising toward the wrong outcome and getting steadily better at it.
When we audit an underperforming ad account, the same problem appears more often than any other. The conversion action being optimised toward is not a sale or a qualified lead. It is a page view, a click on a button, or a form load.
Why this is worse than no tracking at all
Smart bidding is genuinely good at producing whatever you tell it to produce. Tell it to find people who view your contact page and it will find them, cheaply and in volume. None of them will call you.
With no tracking, at least you know you are guessing. With bad tracking you get confident reports about the wrong number.
What to measure instead
Measure the thing closest to revenue that you can reliably observe. For ecommerce that is a purchase with its value attached. For lead generation it is a qualified enquiry, ideally fed back from your CRM rather than counted at the form.
Offline conversion import is underused
If your sales close on the phone weeks later, importing that outcome back into the ad platform is the single highest value thing you can do. It lets the algorithm optimise toward closed business rather than form fills, and it takes an afternoon to set up.
Verify, do not assume
Fire a test conversion. Watch it arrive. Check the value is right and the currency is right. Then check for duplicates, because double counting is almost as damaging as counting the wrong event.
Frequently asked
Compare the platform conversion count against your actual sales or CRM enquiries for the same period. If the numbers are not within a reasonable margin, something is counting the wrong thing.
Sending the real outcome of a lead, whether it closed and for how much, back to the ad platform so bidding optimises toward revenue rather than form submissions.
Want this handled for you?
We build and run the sites we write about. Tell us the target and we will scope it.
Start a project