Same budget,
+21% more media-driven volume
with Marketing Mix Modeling

This engagement is less about a single headline number and more about the analytical work behind it. Five capabilities of the AD cube MMM did the heavy lifting.

1. Telling real drivers from coincidences 

The first job of the model is to separate what really drives sales from what simply moves alongside them. A web-analytics conversion metric looked closely tied to sales, but it was mostly an effect of TV and video-on-demand advertising, not a cause of its own. Treating it as a driver would have made TV and video look weaker than they truly are. For the same reason, the model set aside a few variables that follow sales rather than cause them, so the results reflect what actually moves the business.

2. Counterfactual analysis of key events

This kind of counterfactual analysis was run across a broad set of one-off events and extra variables beyond the media mix, not just these two. As an example: a study supporting the product and the launch of a new variant reshaped the brand’s trajectory. By comparing the sales trend before and after each event, the model isolated two distinct effects: a lasting level shift in volume sustained over the following two years, and a growth shift, with monthly sales rising by roughly +23–27%. This was fed back into the model to make it more robust.

3. Every channel has a saturation point

Two very similar sampling channels were delivering very different returns. The reason was saturation: past a certain point, every extra euro on a channel produces less and less. One channel still had room to grow, while the other was already pushed well beyond its best-performing point. That’s why its return was lower, and why budget should move toward the channel with headroom.

4. What-if scenarios

The model can replay history under different budgets. The brand asked whether TV spend was better concentrated in a burst or distributed over a longer period. At the same budget, the distributed plan outperformed the burst plan by roughly +6.5% on the TV contribution alone, by avoiding the saturated region.

5. Budget optimization

Finally, the AD cube optimizer reallocated the full 2025 budget (unchanged) across channels and time, using the estimated saturation and lag curves. Total sellout rose (+1.76%). Because about 92% of sellout comes from non-media base factors (price, seasonality, brand strength), the real story is the media lever: media-driven volume grew +21.3% on the same spend. Measuring that non-media base explicitly, rather than folding it into the media numbers, is what makes the +21.3% result trustworthy.