Ad Attribution: Stop Letting the Ad Platform Grade Its Own Homework

Last updated: June 2026.

Run ads on both Meta and Google for a month, then add up the conversions each platform reports. You will find you "made" more sales than you actually made. Both platforms claimed the same customers, because both saw a touch, and both are financially motivated to take the credit. The ad platform is a referee that gets paid more when it blows the whistle in its own favor, and we keep asking it for the final score.

Independent ad attribution means measuring ad-driven revenue on your side, with a tool that has no incentive to flatter any channel. It is the only version of ad ROI you can trust, and in agency work it is the single biggest source of "wait, we were losing money on that campaign the whole time" moments.

Why platform-reported ad numbers are structurally inflated

This is not a conspiracy, it is incentives and mechanics:

• Self-attribution bias. Each platform attributes a conversion to itself if it saw any qualifying touch within its (generous, self-defined) window. Two platforms touching the same buyer both claim the full conversion. Sum them and you exceed reality.
• Generous lookback windows. A 7-day-click, 1-day-view window lets a platform claim a sale it barely influenced, because someone glanced at an ad days before buying for unrelated reasons.
• The post-iOS blind spot. Since App Tracking Transparency, the pixel sees less and platforms fill the gap with modeled (estimated) conversions. Modeled data is a forecast wearing a measurement costume, and it conveniently rounds in the platform's favor.

None of this makes the platforms useless. It makes their conversion reports unfit to be your source of truth.

What independent attribution looks like

The principle is simple: one neutral system, your own, watches every channel by the same rules and ties each to actual revenue ↗.

• Every ad click arrives carrying its campaign identity (through tracked URLs and parameters your system reads on landing).
• Your analytics records what that visitor did, on the same footing as organic, email and social, with no channel allowed to grade itself.
• Your payment connection ties the eventual purchase to the channel that genuinely brought the visitor, by one rule, applied to everyone.

Datalenk credits first touch, and there is no model selector, on purpose: a dial you can turn until the numbers flatter the channel you already believe in is not a referee. The whole reason you left the ad platform's report was that it was written by an interested party. Handing you a knob to write your own would just move the problem.

Now Meta's claimed 50 conversions and Google's claimed 40 resolve into your actual 60, attributed by one consistent referee. The gap between "claimed 90" and "real 60" is exactly the budget you were about to waste.

Get the score from a neutral referee. Datalenk measures every channel, including paid, by the same rules and ties each to real revenue, so no platform grades its own homework. 14-day trial, card required, cancel in two clicks. Start your trial.

Use platform data for optimization, your data for truth

The honest workflow is not "ignore the ad platforms". It is a division of labor:

• Inside the ad platform: use its data to optimize creative, audiences and bids. The platform's signal is good for steering, because it is fast and granular.
• In your independent system: judge whether the channel is profitable at all, and how it compares to your free channels. This is the budget decision, and it must come from a source with no skin in the game.

Teams that conflate the two (using the optimizer's self-report as the profit verdict) routinely scale campaigns that lose money, because the referee told them they were winning.

How it fits your ad stack, and the one thing we do not do

You keep running ads where you run them. Datalenk adds the independent measurement layer: every ad click lands carrying its campaign, gets measured on the same footing as every other channel, and ends in a revenue number tied to a payment that actually happened.

Now the honest limit, stated plainly because you will hit it in week one: Datalenk does not pull your ad spend from Meta or Google, so it does not compute ROAS for you. It shows revenue by campaign, never revenue over spend. Building the spend pipe would mean asking you for write-capable access to your ad accounts, and it would put us one API deprecation away from a silently wrong number in the exact place a wrong number costs the most.

So the division of labour is: you bring the spend, we bring the un-inflated revenue. Copy your real spend out of the ad manager (it is the one number the platform has no incentive to overstate), put it next to Datalenk's revenue by campaign, and divide. It takes a minute a week and it is the only ROAS in your stack that was not computed by the party being graded. In practice that minute is what catches the campaign that has been quietly underwater since March.

FAQ

Why do my ad platforms report more conversions than I actually got? Because each platform self-attributes any conversion it saw a touch for, within its own generous window. When several platforms touch the same buyer, each claims the conversion, so the sum exceeds your real total. Independent measurement resolves the double-counting.

What is independent ad attribution? Measuring ad-driven conversions and revenue with your own neutral system rather than trusting each ad platform's self-report. One referee, same rules for every channel, tied to actual revenue.

Should I stop using Meta and Google's reporting? No. Use it to optimize creative and bids, where its speed and granularity help. Do not use it to decide whether a channel is profitable; that verdict needs a source without a financial stake in the answer.

How does iOS affect ad attribution? App Tracking Transparency reduced what the pixel can see, so platforms fill gaps with modeled (estimated) conversions. Independent, first-party measurement on your own site is less affected and is not motivated to estimate in any channel's favor.

Does Datalenk calculate ROAS? No. It does not ingest your ad spend, so it reports revenue by campaign and never a return-on-spend ratio. You bring the spend from the ad manager, Datalenk brings the un-inflated revenue, and you divide. The division takes a minute; getting the numerator honest is the hard part.