Marketing Dashboard FAQ

MMM aims to report the true incrementality of your media spend. Google Analytics reports the attribution of each media - it doesn't tell you how much of that was truly incremental. Therefore, it is expected to have major differences between MMM and Google Analytics reported attribution.

Example: Your Google Search Brand looks to be much more efficient according to Google Analytics - however, not all of your brand searches are truly incremental and driven by media. Many customers knows you already and come to your websites, regardless if they have seen your ad or not.

Sellforte’s Marketing Mix Modeling identifies promotion-driven sales by using the customer’s sales data to detect when promotions have been active and how strong those promotions have been over time. This information is provided to the model as daily promotional activity, allowing it to evaluate how changes in promotional intensity have historically influenced total sales. Based on these patterns, the model separates the share of sales uplift that can be attributed to promotions from the baseline and from other marketing drivers. In the Sellforte UI, promotion-driven sales are shown as their own component, making it easy to understand how promotions have contributed to incremental sales and how their impact has evolved over time.

Unlike attribution tools, which typically struggle to isolate whether campaign performance was driven by the promotional mechanic itself or by the media behind it, MMM can explicitly disentangle these effects. This provides a clearer picture of the true incremental role of promotions versus media, which is especially valuable during high-promotion periods such as Black Week and other seasonal peaks where both media activity and discounting levels intensify simultaneously.

Defining investments for Own Channels, such as email and SMS, is not as straightforward as for paid media channels, since these channels often do not have a direct monetary spend tied to each send. In most implementations, Sellforte uses a proxy approach that reflects the relative scale of Own Channel activity. When data is available from platforms such as GA4 or Klaviyo, we can observe how much sales value these platforms have attributed to email or SMS. A common method is to estimate spend as 10% of this attributed value, providing a practical approximation of the operational and resource investment behind these channels. This gives the model a consistent input that reflects channel intensity over time, even without explicit spend data.

The resulting MMM output can then be interpreted in relation to this proxy investment. For example, if a platform attributes 100 USD in sales to email and the proxy spend is defined as 10% of that amount, the spend estimate is 10 USD. The model then determines that only a portion of the attributed sales, say 50%, was truly incremental. In this example, the incremental sales would be 50 USD, and spend 10 USD, which would give us incremental ROAS of 5.