The essentials
- Definition: attribution is the rule that shares the credit for a conversion between the different touchpoints in a customer journey.
- Models: last click, first click, linear, time decay, position-based, and the data-driven model — the only one available in GA4 since 2023.
- A 2026 fact: the conversions reported by Google Ads, Meta and LinkedIn add up to 30 to 80% more than the real total, each platform claiming the same sales.
- The rule: platform figures are for optimising within a channel; a neutral source (GA4 or your CRM) is for deciding between channels.
Theattribution is the rule that decides which channel, campaign or piece of content gets the credit for a conversion. A customer who discovers a company in a YouTube video, comes back through a Google search, reads an article, receives an email and then buys after clicking a retargeting ad has produced one sale and five touchpoints. Deciding which one "made" the sale is a convention, not a truth: that is what attribution is about, and the leading cause of disagreement over marketing budgets.
The attribution models
| Model | Split | What it favours | When to use it |
|---|---|---|---|
| Last click | 100% to the last touchpoint | Brand search and retargeting | The historic default; simple but misleading |
| First click | 100% to the first touchpoint | Awareness and discovery | To measure what brings people into the journey |
| Linear | An equal share to each touchpoint | Journeys with many channels | Long cycles, an explanatory reading |
| Time decay | More weight to recent touchpoints | The bottom of the funnel | Short cycles |
| Position-based | 40% to the first, 40% to the last, 20% shared | Entry and close | A readable compromise |
| Data-driven | Weights calculated by comparing converting and non-converting journeys | What actually changes the outcome | The only model Google offers in GA4 and Google Ads since 2023 |
The data-driven model has become the reference at Google, which removed the rules-based models from GA4 and Google Ads. Its principle is to compare journeys that convert with those that don't, to estimate each touchpoint's contribution; the detail is set out in data-driven attribution.
Conversion windows
An attribution window defines how long a touchpoint can be credited with a conversion. Google Ads uses 30 days after a click by default, Meta 7 days after a click and 1 day after a simple view, LinkedIn 30 days. Those settings explain part of the gap between platforms: the same sale can be claimed by Meta for a video watched the day before and by Google for a click the previous week. The view-through conversion, counted with no click, is the most debatable: it mechanically inflates retargeting results, as explained in remarketing.
Why the platforms don't add up
Every ad platform measures within its own universe and claims any conversion preceded by one of its touchpoints, with no knowledge of what the others did. Adding up the conversions reported by Google Ads, Meta and LinkedIn gives a total above the real number of sales, by 30 to 80% depending on how many channels are active. On top of that come measurement losses: consent refusals, blocked third-party cookies, device switches. Modelled data fills part of the gap, without allowing any cross-checking between platforms.
How to decide in practice
- Two uses, two sources: the platform's figures are for optimising within a channel (which campaign, which creative); a neutral source is for comparing channels with one another.
- The neutral source: GA4 with data-driven attribution for B2C, your CRM with the original source kept through to signature for B2B; the setup is described in attribution in GA4.
- The overstatement ratio: each month, set the sum of reported conversions against the real total; a ratio above 1.8 signals a channel claiming sales it didn't produce.
- Assisted conversions: also read what your discovery channels bring earlier in the journey, invisible on last click.
- Incrementality tests: for significant budgets, switching a channel off in one area or for a control group measures its real contribution, which no model does.
- Brand kept separate: conversions coming from your company name are counted separately, otherwise they flatter whichever channel captures them.
How GreenRed helps
Rather than juggling several tools, GreenRed's return on investment module brings these metrics together in a single dashboard, compares them over time and tells you which actions come first. You can try it free, with no card, from the Pricing.
Frequently asked questions
What is attribution in marketing?
The rule that shares the credit for a conversion between the touchpoints in the customer journey: ads, search, emails, social media. It is a measurement convention, not an observable truth, and the model you choose changes how each channel's performance reads.
Which attribution model should you choose?
The data-driven model where it is available: it estimates each touchpoint's contribution by comparing converting and non-converting journeys. It is now the only one GA4 and Google Ads offer. Last click remains useful as a simple benchmark, bearing in mind that it overvalues brand and retargeting.
Why do the platforms' conversions add up to more than my sales?
Because each ad platform claims any conversion preceded by one of its touchpoints, with no knowledge of the others. The same sale is therefore counted several times, producing a gap of 30 to 80% depending on how many channels are active. Only a neutral source shares conversions between channels.
How do you measure a channel's real contribution?
With an incrementality test: switch the channel off in a geographic area or for a control group and compare sales with an area where it stays on. It is the only method that measures what the channel adds, and it requires enough budget and volume.