Skip to main content
All CollectionsFAQModeling FAQMarketing Mix Modeling FAQ
How do I interpret marketing mix optimization results?
How do I interpret marketing mix optimization results?
Updated over a week ago

The market mix optimization results page shows three widgets of interest:

  • Marketing spend mix: this shows the average marketing spend mix for (i) the training period, generally the dataset less the last 13 periods which were held back for testing, (ii) the baseline period, which includes the last 13 periods, and (iii) the optimum mix computed during the baseline period.

  • Predicted outcome: this shows the mean predicted KPI for the same three periods as above.

  • Optimization detail: this shows for the KP{I and each media channel the following metrics:

    • Baseline: value during the baseline period. For media spend it is based on actual value. For the outcome it is based on model predictions so the change between baseline and optimum is the change due to media mix optimization only, everything else being held constant.

    • Baseline mix: percent of total during the baseline period

    • Optimized: optimized value during the baseline period

    • Optimized mix: percent of optimized total during the baseline period

    • Variance: absolute change between baseline and optimized scenarios

    • Variance percent: percent change between baseline and optimized scenarios. Will be particularly helpful for the KPI (dependent variable)

    • Mix Shift: percentage point change in mix between baseline and optimized scenario

Did this answer your question?