The baseline in a marketing mix model is the level of sales you would achieve with no advertising at all. It includes the effect of brand equity, distribution, pricing, seasonality, and organic demand. Everything above the baseline is the contribution from your marketing activity.
When an MMM provider shares your results, one of the first charts they show is a waterfall or stacked bar that breaks total sales into baseline and incremental contributions by channel. Understanding what baseline means, and whether it is set correctly, is one of the most important checks you can do on any model.
Why baseline matters so much
The baseline determines how much credit marketing gets. If the baseline is set too high, your media looks less effective because most of the sales are assigned to "brand and organic" rather than to your channels. If it is set too low, your media looks more effective than it really is.
Getting baseline right is therefore one of the most consequential modelling decisions. It affects every ROI estimate in the model and every budget recommendation that follows.
What drives the baseline
The baseline is not a single thing. It is the combined effect of several factors that drive sales regardless of media spend in a given week.
- Brand equity: the accumulated recognition and preference that customers already have for your brand
- Distribution: how widely your product is available in stores or online
- Price: your standard selling price relative to competitors
- Seasonality: predictable demand patterns tied to the time of year
- Long-run advertising effects: the slow-building brand awareness from years of past campaigns
- Macroeconomic conditions: consumer confidence and spending power
How the model estimates baseline
The model estimates baseline by removing the effect of all the media variables and looking at what remains. In practice, the baseline is the intercept of the regression equation, adjusted for all the non-media factors included in the model. The more completely the model accounts for seasonality, price and external factors, the more accurate the baseline will be.
A common benchmark is that baseline accounts for 40 to 60 percent of total sales for established FMCG (fast-moving consumer goods) brands. For newer brands with heavy media dependence, it may be lower. For very well-established brands with minimal advertising, it can exceed 70 percent. If your baseline looks implausible given your brand's history, challenge it.
The long-run base contribution from advertising
An important nuance is that past advertising contributes to the baseline. A TV campaign you ran two years ago built brand awareness that still drives purchases today, even though the spend has long since stopped. This effect is sometimes called the long-run brand contribution or brand equity build. Capturing it correctly requires a model that separates short-run media effects from their long-run impact on the base.
If your model ignores this, it will understate the long-term value of brand advertising relative to performance channels, which tend to drive short-term, easily measurable effects. This is one reason why some brands who rely only on short-term measurement end up systematically underinvesting in brand.
Questions to ask about your baseline estimate
Ask your modelling partner to walk you through what is driving your baseline and whether it has changed over time. A declining baseline can be an early warning that brand equity is eroding. A rising baseline might reflect successful distribution growth or a long-run payoff from sustained brand investment.
Should we try to grow our baseline?
Yes. A higher baseline means more of your revenue is protected from the short-term fluctuations of your media plan. Sustained brand advertising, improved distribution, and pricing discipline all build the baseline over time.
If our baseline is very high, does that mean our advertising is not working?
Not necessarily. A high baseline in an established brand often reflects the accumulated effect of years of successful advertising. It means the brand has strong organic demand. The question to ask is whether incremental media spend is generating a good return on top of that base.
How do we know if the baseline is set at the right level?
One check is to look at what the model predicts your sales would be during a period when you ran no advertising. If you have ever gone dark on all media for a period, you can compare the model's baseline prediction to actual sales in that period. That is one of the most direct validations available.