Multi-touch attribution (MTA) tracks individual customer journeys across digital touchpoints. Marketing Mix Modelling (MMM) measures the aggregate contribution of all media channels over time. They answer related but different questions, and they often disagree with each other.
If you have used both tools, you have probably noticed that they give you different answers about which channels are working. That disagreement is not a problem with either tool. It reflects the fact that they are measuring different things.
Understanding the difference helps you use both more effectively, and helps you know which one to trust when they contradict each other.
What is multi-touch attribution?
Multi-touch attribution tracks the digital touchpoints a customer encounters before making a purchase: a display ad on Monday, a social media post on Wednesday, a branded search on Friday, a purchase on Saturday. Then it assigns credit to each of those touchpoints according to a chosen rule.
Different MTA models assign that credit differently. Last-click gives 100% of the credit to the final touchpoint before purchase. First-click gives 100% to the first touchpoint. Linear splits credit equally across all touchpoints. Data-driven models use machine learning to assign credit based on which touchpoints statistically correlate with conversion.
What is Marketing Mix Modelling?
Marketing Mix Modelling does not track individuals at all. Instead, it looks at aggregate patterns over time. It takes your total weekly or monthly spend across each channel alongside your total sales, and uses statistical modelling to estimate how much each channel contributed.
Because it works with aggregate data, MMM can measure channels that MTA cannot see, including TV, radio, out-of-home advertising, and any other channel that does not leave a digital tracking cookie.
Where each one works well
Where MTA has the advantage
- Granular, campaign-level optimisation decisions made daily or weekly
- Understanding the sequence of digital touchpoints before a conversion
- Comparing creative performance within a single platform
- Real-time feedback on digital campaign performance
Where MMM has the advantage
- Full-funnel measurement including offline channels like TV and radio
- Separating marketing impact from external factors like seasonality and competitor activity
- Strategic budget allocation across the entire media mix
- Long-term brand-building effect measurement
- Working in a privacy-first environment without relying on cookies or user-level tracking
Why they disagree
If you put MTA and MMM results side by side, you will almost always see disagreements. MMM typically attributes more revenue to brand and upper-funnel channels like TV. MTA typically gives more credit to lower-funnel channels like branded search and retargeting, because those are the last touchpoints before a purchase.
This makes sense when you think about what each is measuring. Brand activity builds the awareness and intent that eventually leads someone to search for you. But MTA cannot see that TV ad, so it gives all the credit to the search click that happened at the end of the journey.
A useful mental model: MTA tells you who was in the room when the customer decided to buy. MMM tells you what actually persuaded them to buy in the first place.
The privacy problem with MTA
Multi-touch attribution relies on tracking individual users across the web, typically using cookies or device identifiers. As browsers have restricted third-party cookies and users have increasingly opted out of tracking, MTA has become less accurate.
The proportion of customer journeys that can be tracked end-to-end has fallen significantly. In some markets and for some audiences, the majority of conversions are no longer trackable at the individual level. MTA built on incomplete data produces misleading results.
MMM is not affected by this problem. It uses aggregate data, so privacy changes do not reduce its accuracy.
Which one should you use?
The honest answer is that you ideally want both, used for the decisions they are each suited to. Use MTA for in-flight campaign optimisation within digital channels. Use MMM for strategic budget decisions across your full media mix.
When the two disagree on a channel's value, treat that as a question to investigate, not a problem to ignore. Running an incrementality test on the disputed channel is often the fastest way to resolve the disagreement with real evidence.
Is MTA dead?
Not dead, but significantly diminished by privacy changes. It still has value for granular digital campaign optimisation within platforms that have first-party data. For strategic budget decisions, it is no longer the right primary tool.
Can MMM replace MTA entirely?
MMM is better for strategic decisions, but it is not granular enough for daily campaign optimisation. Most marketing teams benefit from using both, with clear boundaries around which decisions each informs.
Why does my MTA show higher performance for paid social than my MMM?
This is very common. MTA picks up the last-click conversions from people who were already likely to buy. MMM estimates only the revenue that would not have existed without the paid social campaign. The MTA number includes both incremental and non-incremental conversions.