Saturation in MMM refers to the diminishing returns you get as you spend more on a channel. Every channel has a point at which additional spend produces progressively smaller sales uplifts. The model captures this with a curve that flattens at higher spend levels.
If doubling your TV budget doubled your sales, there would be no optimal budget. You would just keep spending. The reason that does not work is saturation: each additional pound reaches fewer new people, or repeats exposure to people who have already been persuaded. The model quantifies exactly where this flattening happens for your brand.
The shape of the saturation curve
The saturation effect is modelled using a curve, most commonly a Hill function or a logistic curve. Both produce the same general shape: steep gains at low spend levels, then a gradual flattening as spend increases. The curve never turns negative in a well-specified model; it just approaches a ceiling.
The steepness of the curve, and the point at which it starts to flatten, varies by channel and by brand. A niche brand with a small addressable audience will saturate faster on a given channel than a mass-market brand with broad appeal.
Why saturation matters for budget decisions
Saturation tells you where your marginal return on investment is highest. The marginal return is the additional revenue generated by one more pound of spend. At low spend levels the marginal return is high. As you approach saturation it falls. The optimal budget point is where the marginal return equals the cost of the next pound spent.
- Channels operating below saturation: adding budget produces good returns, underspending likely
- Channels operating at saturation: returns are diminishing, reallocation to other channels may improve efficiency
- Channels operating above saturation: each additional pound may be producing less than a pound of revenue, significant overspend
Most brands discover through their first MMM that at least one channel is operating well into the saturation zone. The most common culprit is paid search brand terms, where you are effectively paying to capture demand that would have arrived anyway.
The relationship between saturation and ROI
Average ROI (total revenue divided by total spend) and marginal ROI (the return on the next pound) are different numbers, and saturation is why. A channel can have a strong average ROI of 3.0 while the marginal ROI has fallen to 0.8. You are still making money overall, but the last pounds spent on that channel are losing money. The model helps you see both numbers.
Budget optimisation uses marginal ROI to find the reallocation that produces the best overall return. It shifts spend away from channels where the marginal return is low and toward channels where it is still high.
Saturation and the optimal budget level for your brand
The saturation curve also tells you something about your total budget level, not just how to allocate it. If all your channels are well below saturation, the model is suggesting you are underinvesting overall and could deploy more budget profitably. If they are all at or near saturation, the model is suggesting there is limited room to grow efficiently without entering new channels or markets.
Does saturation mean we should stop advertising on a channel entirely?
Not necessarily. Even at high spend levels, a channel may still be generating a positive return overall, just a smaller one per additional pound. The decision to cut or pause a channel depends on whether the budget saved can generate a better return elsewhere.
How do we know if our model is accurately capturing the saturation curve?
One check is to look at whether the model-implied optimal spend for each channel is within the range you have actually tested. If the model suggests doubling your current TV spend, but you have never run at that level, treat the prediction with caution. Models are more reliable within the spend range they have observed.
Can saturation change over time?
Yes. As your brand grows, your addressable audience expands and saturation may arrive later. Changes in media costs, competitive activity, and platform algorithm changes can all shift the point at which a channel saturates. This is one reason why updating your model annually is worthwhile.