The 4 conversion-tracking mistakes I see in every new account audit

Every optimization decision in a paid account rests on conversion data. Every bid, every budget shift, every "this campaign works" call. When the data is wrong, everything built on it is wrong. These are the four mistakes that show up in almost every audit.

1. Counting actions that aren't outcomes

Page views, scroll depth, button clicks, "engaged sessions." None of these pay the bills, and none of them belong in your primary conversion column. When they're in there, the algorithm optimizes toward low-cost engagement instead of customers, and the account looks better as it performs worse.

2. Double counting

A form submit fires a conversion. The thank-you page fires another. Now every lead counts twice, your cost-per-lead looks half its real size, and every scaling decision is based on numbers that are 100% inflated. Pick one source of truth, fire the conversion exactly once, and make it the thank-you page. That's the cleanest.

3. No value attached

A lead for a $200 service and a lead for a $20,000 job count the same in an account with no conversion values. The algorithm can't tell them apart, so it optimizes for volume, and volume skews toward the smallest jobs. Even rough values help. Assign your average job value by form type, and bidding chases revenue instead of raw lead count.

4. The pixel that broke silently

Tracking isn't set-and-forget. Site redesigns remove tags, consent banners block scripts, iOS updates strip signals. The account keeps spending. The data quietly stops arriving. The algorithm reverts to guessing. If your conversion count dropped and your traffic didn't, the pixel broke. Check it before you touch a single bid.

The uncomfortable summary

Most "the ads stopped working" stories are measurement stories. It's also the first thing checked in every audit I run, because until the tracking is right, no other opinion about the account means anything.