Senior stakeholders do not want to hear about model fit or adstock parameters. They want to know where their money is going, which channels are working, and what you recommend doing differently. Lead with the business implication, then show the evidence.
MMM readouts are full of statistical concepts that make perfect sense to the analysts who built the model and almost no sense to a finance director sitting through a 45-minute presentation. The way you frame the results determines whether they get acted on or filed away.
Start with the answer, not the methodology
Open with a single slide that states the three to five most important findings in plain language. "We are overspending on paid search brand terms by an estimated 30 percent. Reallocating that budget to connected TV could generate an additional X in revenue with the same total spend." That is a lead. The methodology slide comes later, or not at all depending on your audience.
CFOs respond to return on investment numbers and budget scenarios. CEOs respond to competitive positioning and growth opportunity. Tailor the opening minute of your presentation to the priorities of the person in the room.
The four charts that tell the story
- Revenue decomposition: a waterfall or stacked bar showing how much of total sales came from baseline, each media channel, and other factors such as price and promotions
- ROI by channel: a bar chart of return on investment for each channel, making it easy to rank efficiency
- Saturation curves: a chart showing where each channel sits on its spend-vs-response curve, indicating which are under- or over-invested
- Budget scenario comparison: a before-and-after showing current allocation vs. recommended allocation, with the projected revenue difference
Handling uncertainty in the room
A CFO will ask how confident you are in the numbers. Do not say "very confident." Say "the model estimates TV ROI between 1.4 and 2.9 with 90 percent confidence." Then explain what that means: even at the low end of the range, TV is generating a positive return, so the directional recommendation holds.
Presenting ranges rather than single numbers takes courage, but it builds credibility. A finance director who later discovers that a single-point estimate was masking significant uncertainty will trust future presentations less. Honest uncertainty is a feature, not a weakness.
Anticipating the questions you will get
Prepare for these three questions in every senior MMM presentation. First: "How do we know this is right?" Have the validation evidence ready, including out-of-sample accuracy and any experimental calibration. Second: "What would happen if we just did nothing?" Show the baseline trajectory. Third: "How quickly will we see the results of any changes?" Have a realistic view on timing, accounting for adstock and the lag between budget changes and measurable sales impact.
Turning the presentation into a decision
End every MMM presentation with a clear, time-bound recommendation. Not "we should consider reallocating budget" but "we recommend moving 15 percent of Q3 paid search spend to connected TV, with a review after 12 weeks." Attach a number to the expected benefit and a name to who owns the change. That is how a presentation becomes an action.
How do we handle stakeholders who do not trust the model?
Show them the validation evidence first. If the model accurately predicted sales in a period it was not trained on, that is compelling. If they remain sceptical, propose a small budget test in the direction the model recommends. Real-world results are more persuasive than any slide.
Should we share the full technical model output with senior stakeholders?
No. The technical output is for the analytical team. Senior stakeholders need a business summary with three to five clear findings, the supporting evidence, and a recommendation. Offer to share the technical appendix for anyone who wants to go deeper.
How do we present MMM results alongside last-click attribution data, which often tells a different story?
Acknowledge the difference directly. Explain that last-click attribution gives 100 percent of the credit to the final digital touchpoint and misses offline channels entirely. MMM looks at the full picture. For channels like TV, MMM is more reliable. For short-term digital optimisation, both data sources have value.