Google, Meta and your affiliate network can each claim the same sale, so your dashboards add up to more sales than you made. A shared measurement standard can make those numbers comparable, but only you can choose the single source of truth every channel is judged against: one revenue or margin figure from your own finance system or CRM, not an ad account.
On 20 October 2026 the IAB opened a 15-day public comment, running to 4 November 2026, on a guide meant to make every advertising channel comparable on one page. A shared method for reading Google, Meta, LinkedIn and the rest side by side is worth having.
The guide agrees the method and leaves the goal to you
The guide is an attempt to agree how channels are compared: which attribution model you select, how a sale two platforms both claim is deduplicated, and how a customer is matched across devices and logins (identity resolution). That is useful housekeeping. It settles the method. It does not settle the goal, which is what gives a comparison any meaning.
Each platform books the sale by its own rules
Google Analytics 4 credits a sale through data-driven attribution; Meta claims it through its Conversions API; an affiliate network settles on the last click before purchase. Say a £1,000 sale: it can surface as three conversions across three dashboards. Meta's own deduplication documentation shows how easily it happens inside a single platform: without it, the browser pixel and the server both record the same purchase, so one sale reads as two. Deduplicating that arithmetic is the easy half of the problem, and it is the half a shared method can fix.
A platform's dashboard reports its own best case
A platform's dashboard is the platform's best case for its own spend. Some of the figures are modelled conversions, estimates the platform fills in where consent controls like Google Consent Mode stop it seeing every action. None of it deducts the sale that would have happened anyway. That gap has a name: incrementality, or lift, the extra business a channel actually caused. It is the hard half of the problem.
The standardisation camp is half right
Performance agencies and the platforms themselves sell the opposite view: that unified measurement, a shared standard plus data-driven attribution and marketing-mix modelling tools like Meta's open-source Robyn, finally makes channels comparable. They are half right: standardised method removes the arithmetic mess. It does not remove the self-grading bias, because every input is still a number the seller produced, and a tidier comparison of biased inputs is still biased. A cross-channel standard would outrank your own books only if it made platforms report independently audited, deduplicated incremental outcomes against one agreed definition. Until a standard requires that, your own books stay the better judge.
Make your own books the single source of truth, then test each channel's lift against it
Pick one outcome, make it the single source of truth every channel is judged against, then split the credit by testing what each channel actually causes. Take the figure from a system you own, not one a platform hands you:
- Revenue booked in your finance system: money received, not sessions attributed in an ad account.
- A lead marked won in your CRM: a human judgement, not a form fill a platform claims credit for.
- A margin figure you would defend to your board: reconcilable to cash, not a conversion count no one can tie back to it.
Rank channels on last-touch booked revenue alone and you will defund the awareness that feeds it, the upper-funnel channel that looks worthless on the final click but can be what creates the demand in the first place. So the booked figure is the denominator, and credit is split by testing incrementality: hold a channel out in some regions (a geo holdout) and measure how far their booked total moves against matched regions where it keeps running. A channel whose absence drops your revenue earned its place, whatever its last-click record shows. A channel you can switch off with no measurable effect was being paid for demand it never created. The only numbers you can fully trust are your own.
Settle your own yardstick before the guide is final
Comment on the guide if your sector depends on it; the window is open to 4 November 2026. But treat the consultation and your own measurement as two separate jobs. One shapes an industry convention that may arrive slowly, if at all. The other decides, this quarter, whether next year's board pack ranks your channels by what they cost or by what they returned. Start this quarter, with your highest-spending channel. Do that, and any standard, whenever it lands, reads as confirmation rather than correction. That is the data and automation work worth doing now, because it pays whenever the standard arrives, or if it never does.
