What is multi-touch attribution?
It shares the credit for a sale across every ad and visit that led to it, instead of handing all of it to the last click.
45 minutes on relevant demo data. No preparation needed.
Five touches, one order
Three mistakes that move budget the wrong way
Each looks reasonable on its own, and each takes budget away from the ads that start journeys.
Mistake 1: giving the last click everything
Last-click ignores the whole journey, including the ad that introduced the brand. Without that first touch, the sale might never have happened.
- Overpays brand search and retargeting, which close demand others created.
- Underpays video and social, which start journeys and rarely get the last click.
Mistake 2: adding up the platform reports
Each platform grades its own homework. Every one that touched an order can claim all of it, so the reports added together can exceed what you sold, while the ads that started the journey can still get nothing.
- Different windows Google Ads counts clicks for 30 days, Meta counts views for one.
- No referee nobody deduplicates across platforms.
Mistake 3: leaving out the ads people only saw
Display and video mostly work without a click. A tool that sees only website visits analyses half a journey and calls it the whole story.
- Walled gardens Meta, TikTok and YouTube do not export impression-level data.
- Synthetic Impressions reconstruct those views so they can earn credit.
One sees too little. The other claims too much.
Roivenue models in the Meta views GA4 misses, then splits each sale across the channels that earned it.
The easy alternatives, and why they fall short
Every model splits the same order. Only the rule changes.
Same order, six answers
Last-click gives everything to the brand search. First-click swings to the other extreme. Linear, time decay and U-shaped spread the €240 by a rule chosen in advance, so they are easy to explain and still a guess: a fixed rule cannot tell a scrolled-past ad from a comparison-site click.
- Rule-based models encode an assumption.
- Data-driven models learn the weights from your own journeys, if the data underneath is complete.
How Roivenue does it
Even the best model needs complete data. So we collect first, join second and credit last.
Our own pixel records visits and ad views; Synthetic Impressions model, from connector data, the views Meta, TikTok and Snap keep to themselves.
Path Stitching connects the phone that viewed the ad and the laptop that bought, without third-party cookies.
An AI model trained on your converting and non-converting journeys decides what each touchpoint was worth. Then the Budget Optimizer shows where the next euro pays back.
Questions about multi-touch attribution
Last-click gives 100% of the credit to the final touchpoint before the purchase. Multi-touch attribution shares it across every interaction in the journey, from the first ad seen to the last click. Last-click systematically underpays paid social, display and video, and overpays branded search, which often captures demand created elsewhere.
No. Multi-channel describes the scope: the journey crosses several media. Multi-touch describes the method: credit is shared across interactions. A multi-channel report can still be last-click.
It depends on how many channels you run and how long the journeys are. As a rule of thumb, 100 to 200 conversions a month is a reasonable minimum for a straightforward setup. Below that, a rule-based model is usually the more honest choice.
Yes, when it is built on first-party data and identity resolution. Models that relied on third-party cookies to join devices have degraded; models built on first-party measurement have held up.
No. Attribution follows individual journeys and updates daily; marketing mix modelling estimates channel contribution from aggregate history, offline included. They answer different questions and work best together. Roivenue does attribution; ScanmarQED, our parent company, builds MMM.
See it on relevant demo data.
45 minutes with a consultant. You leave with a summary for Finance and your price.