Your own data flags wasted ad spend first

Two numbers you already hold surface the problem before any tool does, and neither requires anyone to accept a methodology.

The two

The claim ratio

Sum what every ad platform says it caused for a single window. Divide by what your store actually took in the same window. Above one is expected, because each platform counts conversions it touched and touches overlap, but the size tells you how much of your reported performance is double counting.

Twenty minutes, no tooling, and nobody has to agree on anything methodological, because it is arithmetic on two totals. Method in the claim ratio.

Coverage

The share of your orders that arrive with a usable source. The remainder does not vanish; it typically gets credited to direct, which flatters direct and starves whatever actually caused the visit. A falling coverage figure means an increasing share of your budget is being allocated on inference. The mechanism is in the unassigned traffic problem.

Why these come first

Because they are yours. Your order table is transactional and cannot be revised by a third party, and your analytics account is an asset you keep regardless of which measurement tool you use. Any window you can still export is a window any method can read, now or in three years. The general argument is in you do not own your attribution data.

What they do not tell you

Which channel to cut. Both numbers describe the state of your reporting rather than the performance of any particular channel. They tell you the problem exists and how big it is; they do not locate it.

Locating it needs a per-channel causal read, which is a €99 one-time upload of a Google Analytics export returning an estimate, a confidence interval, a coverage share and a design label per channel, refundable within 30 days, no questions asked.

The order that works

Claim ratio and coverage first, because they are free and they establish the size. Then the read, to locate. Then a geo holdout on the one channel where the stakes justify it. Each step is more expensive and more conclusive than the last.

The habit worth keeping

Recompute both quarterly and note them with the date. The level is interesting once; the trend catches platform rule changes and collection breakages that nothing in your dashboards announces.

Related answers

• Telling the team a channel is wasting spend
• What looks like wasted ad spend but is not
• What instant channel diagnosis can and cannot do
• Own the pipe: attribution API keys and exports
• Your own numbers settle the reported ROAS argument