How to stop attribution debt accumulating

Paying the debt down once is a project.

The four habits

The claim ratio as a vital sign

Sum platform-claimed revenue for a window, divide by actual revenue. Track the number quarterly. The level is interesting once; the trend is what matters, because a widening ratio means the overlap in your reporting is growing and your thresholds are drifting further from reality.

It also catches platform rule changes, which move the ratio without anything in your business changing. Method in the claim ratio.

Coverage as the second vital sign

Coverage is the share of your real orders a read can explain. It degrades slowly through consent refusal, browser policy and tracking loss, and nothing in your reporting announces it. A falling coverage figure means an increasing share of your business is being explained by inference rather than observation.

Dating definition changes

Channel grouping, lookback window, timezone, what counts as an order. When any of these changes, write the date. Without that record, a discontinuity in your series is indistinguishable from a change in your business, and both look like a finding.

This is the cheapest habit on the list and the one that prevents the most wasted investigation.

One deliberate change per cycle

The debt grows fastest in businesses whose spend pattern is pure reaction, because reactive change destroys the comparisons a causal read needs. One change per cycle, held for a full window, with the prediction written down first, keeps the data interpretable. The mechanics are in the attribution loop and the failure mode in the cost of reacting to attribution noise.

The one-line version

Two vital signs, one dated document, one change at a time.

Related answers

• Attribution debt sits with you, not the vendor
• Attribution debt compounds fastest before peak
• What unpaid attribution debt costs a DTC brand
• The attribution loop: change spend, then re-read