When Microsoft Ads silently drops your conversions.

A tracking setup can be correct end to end and still report zero. This is one of the strangest cases we debug, and the checklist it should leave you with.

The symptom.

The UET tag is on every page. The network tab shows the beacon leaving with a clean response. The UET Tag Helper is green, the goal is configured, consent is granted. And the dashboard shows nothing: zero conversions, audiences that never fill.

uet beaconfires, clean response
tag helperno errors
goal setupcorrect
dashboard0 conversions

The cause is not in your code.

Microsoft classifies every advertiser account into an industry. For categories it treats as sensitive, health and wellness for example, UET data is filtered on Microsoft’s side under its privacy rules. Silently. No warning in the UI, no error in the browser, no rejected request to find. The events arrive and are discarded.

If your account landed in a restricted vertical, by accident or by overly cautious classification, your measurement is empty no matter how good the implementation is. The fix is organizational, not technical: get the classification reviewed through Microsoft support. We have seen teams rebuild a working tag three times before anyone thought to ask.

The checklist for every empty-dashboard morning.

The real lesson is the debugging order. Most teams loop forever on the browser side, because that is the part they can see. Put destination-side checks on the list from day one:

1 · browseris the tag present and firing?
2 · consentdid your test session actually grant it?
3 · wiringdo tag id and goal configuration match?
4 · patienceprocessing lag, date range, attribution window
5 · destinationsilent platform-side filters: industry, policy, data-use restrictions

Step 5 exists on every platform, not just Microsoft: ad platforms increasingly filter data by category, region and policy without telling you. A release is only verified when the number lands in the platform, which is why destination-side checks are part of every QA protocol we write.

Dashboard empty while every beacon fires? Book a call, this one is usually quick to diagnose.