A broken conversion tag produces one of the more disruptive failures in digital advertising, because the damage extends well beyond the reporting. Recovery follows a recognisable pattern.
The reporting fails and the bidding follows
Automated bidding depends on a stream of conversion events to estimate the value of each impression, and it adjusts continuously against what it receives.
When events stop arriving, the system does not recognise a fault. It concludes that performance has collapsed and reduces bidding accordingly.
Delivery therefore falls alongside the reported conversions, which means the campaign is genuinely damaged rather than merely mismeasured.
Detection is usually slower than it should be
A partial break affecting one page type or one browser produces a decline that resembles ordinary seasonal weakness.
Teams investigating a downturn tend to examine audience, creative and competition before they examine whether the measurement is intact.
Automated alerting on conversion volume relative to sessions catches this far earlier, and it is a small piece of monitoring that few accounts have configured.
The repair is the easy part
Restoring the tag itself is usually straightforward once the cause is identified, whether it was a site release, a consent change or a container edit.
What cannot be restored is the history, since the events that were never recorded do not appear retrospectively.
The bidding model consequently carries a period of learning built on data that understates reality, and it will continue applying that until new data outweighs it.
Recovery requires patience the situation discourages
The instinct after a fix is to intervene aggressively, raising budgets and adjusting targets to recover the lost period.
Substantial edits reset the learning process, which extends the unstable period rather than shortening it.
The more reliable sequence is to restore the tracking, leave the structure untouched, and allow the system to accumulate accurate events for a defined period before making any judgement.
The reporting gap needs annotating
A break leaves a hole in the historical record that will be compared against in future periods, producing misleading year-on-year figures long afterwards.
Annotating the affected dates in reporting prevents the artefact from being read as a genuine result by someone reviewing it later.
Organisations that skip this step regularly find themselves explaining an inexplicable performance jump a year later, caused entirely by the depressed baseline the break created.