Pay down attribution debt in one quarter
Four moves, roughly one a month, that take a brand from unmeasured allocation to a documented per-channel read with thresholds the whole team agreed.
Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free
By Joris van Huët, Founder & CEOUpdated 3 min read
Four moves, roughly one a month, take a brand from allocation nobody can check to a documented per-channel read with agreed thresholds. The order matters because each one makes the next possible.
The quarter
| Month | Move | Output |
|---|---|---|
| 0 | Measure the balance | A claim ratio and a coverage figure |
| 1 | Get one read you would defend | A per-channel estimate |
| 2 | Reset the thresholds explicitly | Agreed numbers, written down and dated |
| 3 | Test the largest contested channel | One holdout result |
Month zero: measure the balance
Sum what every ad platform claims for one window and divide by what your store actually took. Note your coverage: the share of orders your analytics can assign to any source. The method is in the claim ratio.
Write both down with the date. This is your opening balance.
Month one: one defensible read
Export a window with real spend variation in it and run a causal read. The €99 one-time read on a Google Analytics export shows what each channel caused next to what last-click gave it, with a data-health score rather than a confidence interval, refundable within 30 days, no questions asked.
Do not act on it yet. The output of this month is understanding, and the test of understanding is whether you can name the widest interval without looking.
Month two: reset the thresholds
Old thresholds were calibrated against inflated numbers, so applying them to causal estimates makes every channel look worse overnight.
Agree new ones explicitly, expressed as conditions on the interval rather than the midpoint: upper bound below break-even is a cut signal, lower bound above is a scaling signal, straddling is a third state.
Month three: test one channel
Pick the channel where the read and the platform reporting disagree most and where the budget is largest. Run a geo holdout on that one. One result, properly obtained, does more for the credibility of the whole programme than six months of reports.
What you have at the end
An opening balance, a per-channel read you understand, thresholds nobody can dispute the origin of, and one experimental result. That is a paid-down position, and it took four moves.
What not to attempt in the same quarter
A tooling migration, a warehouse project, or a change in channel grouping. Each of those makes the before and after incomparable, which is the one thing that would undo the work. The safe migration order, if you need one later, is in replacing multi-touch attribution in four steps.
Related answers
Frequently asked questions
How long does it take to fix attribution debt?
A quarter for the core of it: measure the balance, get one defensible read, reset thresholds explicitly, and test the largest contested channel. Each step makes the next possible, so the order matters more than the speed.Why reset thresholds before acting on new numbers?
Because existing thresholds were calibrated against inflated figures. Applying them to causal estimates makes every channel appear to fail at once, which discredits the read rather than the old threshold.What should I avoid during the same quarter?
A tooling migration, a warehouse project, or a change to channel grouping. Each makes before and after incomparable, which undoes the entire point of establishing an opening balance.
Go deeper: Causal attribution, explained.
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.
Keep reading
Terms in this article
- AnalyticsAnalytics is the systematic computational analysis of data. It reveals customer behavior and measures campaign performance.
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Attribution DebtAttribution debt is the gap between what your ad platforms claim drove revenue and what actually caused it, carried quarter after quarter into the budget. It is how marketing debt accrues: allocate on claimed conversions long enough and the plan itself becomes the liability.
- Attribution ReportAttribution Report shows which touchpoints or channels receive credit for a conversion. It identifies which campaigns drive desired actions.
- ClickClick is the action a user takes to interact with a digital advertisement, redirecting them to a website or landing page. Clicks are a fundamental metric for measuring ad engagement and a primary input for click-based attribution models.
- Confidence IntervalConfidence Interval is a statistical range of values that likely contains the true value of a metric. In marketing analytics, it quantifies uncertainty around estimates, indicating the precision of an outcome or causal effect.
- Google AnalyticsGoogle Analytics is a web analytics service that tracks and reports website traffic.
- Multi-Touch AttributionMulti-Touch Attribution assigns credit to multiple marketing touchpoints across the customer journey. It provides a complete view of channel impact on conversions.