tl;dr
MMM is a Klar unique attribution model is an enhancement of the Data-Driven Attribution that goes beyond click-based data
It uses any other signals and correlations between the channels and runs it through machine learning models to redistribute direct & branded traffic to the original channels
This way MMM also takes into account things like word of mouth and its impact on a conversion
You'll find it in our attribution reports next to the other attribution models to be displayed
We are not fans of using static attribution models like First-Touch or U-Shape. Every single user journey is unique and should be treated as such.
Instead, we like to think of attribution like a goal is being scored in football. Not every single person touching the ball is creating equal value.
The Data-Driven models already factors in the impact a touchpoint realistically had on a conversion. But the thing is:
You do not need to have a touchpoint to have an impact on the conversion. 🤯
There are marketing channels and things like Word of Mouth that create awareness for your product without generating a touchpoint.
But are still essential for the conversion to take place.
That is a problem that we now solve - with our Marketing Mix Model (MMM) that has been built on top of our Data-Driven Attribution.
It is ingesting millions of different signals and data points and running it through machine learning model to re-allocate e.g. direct traffic conversions from new customers to the channels that originally created that demand
Our Klar Data-Driven Attribution already factors in the impact a touchpoint realistically had on a conversion. But the thing is:
You do not need to have a touchpoint to have an impact on the conversion. 🤯
There are marketing channels and things like Word of Mouth that create awareness for your product without generating a touchpoint.
But are still essential for the conversion to take place.
That is a problem that we now solve - with our Marketing Mix Model (MMM) that we added to our Data-Driven Attribution.
It is ingesting millions of different signals and data points and running it through machine learning statistical models to re-allocate e.g. direct traffic conversions to the channels that originally impacted it.
Let me explain you the whole thing a bit more vividly and how it relates to the football goal from above in this video: 👇
So for example you might have a customer journey that, based on clicks only, looks like this. But for sure this cannot be the only channel having impacted that conversion. The customer will always know you from somewhere.
So to understand where this conversion is coming from, we look at the impact of other signals like the spend impact from other channels on the branded traffic and conversions, the impressions impact, the engagement impact and post purchase survey signals.
Letting this run through machine learning algorithms lets us derive the channels that probably impacted the conversion in a certain way and lets us re-distribute branded traffic to the original sources. This also includes adding something like Word of Mouth, that is always happening for every brand in a certain percentage and cannot be quantified and taken into account otherwise:
Comparing static last click attribution with our dynamic MMM attribution, MMM is giving you a way clearer and more realistic picture of which channel is actually driving revenue, that goes beyond click-based data.
More on this in the video.
Attribution Models in Comparison
If you want to know more about the different static & dynamic attribution models you find in Klar, how they are to be distinguished and when to use which one, check out this video.