Reading The Signals: ROI, Response Curves, and Synergy. 

An MMM model produces three analytical outputs that together answer the question every marketing director is really asking: where should my money go? Here’s how to read each one. 

Thought Leadership  -  8 min read  -  MASS Analytics 

Once you have a validated marketing mix model, you have something genuinely powerful: an accurate, evidence-based description of how your marketing is performing. But a description is only useful if you know how to read it. 

The model’s most important outputs fall into three categories. Each answers a different question. Together, they form the basis for every significant budget decision a retailer should be making. 

01 – ROI by channel 

What return is each channel generating per pound spent, and how is that changing over time? 

02 – Response curves 

For each channel, where are you on the spend-to-revenue curve and are you approaching diminishing returns? 

03 – Synergy effects 

How do your channels interact, and what is the uplift from running them together versus in isolation? 

ROI by Channel: The Historical Picture

The first thing a model gives you is a clear, comparable measure of what each media channel returned on your investment, not reported by that channel’s own platform, but measured independently, against your actual sales data, accounting for everything else that was happening at the same time. 

This matters more than it might initially seem. Platform-reported ROI is almost always overstated, because platforms claim credit for sales that would have happened anyway, and they can’t see the contribution of other channels happening simultaneously. An independent model produces a different, more accurate figure and the difference between the two is often significant. 

What you get is a year-on-year view of return per pound spent, by channel. The trends within it are as important as the absolute numbers: 

The numbers above are illustrative, but the patterns they show are typical. Some channels are improving in efficiency year on year; others are declining. Some are clearly saturated. Others have room to grow. Without this view, budget decisions get made based on habit, negotiation leverage with agencies, or which channel’s representatives made the most compelling presentation last quarter. 

Response Curves: The Forward-looking Tool

ROI tells you what happened historically. Response curves tell you what will happen next and where the opportunity lies. 

A response curve plots spend on the horizontal axis against revenue generated on the vertical. For every channel in your mix, the model produces a curve showing the relationship between those two numbers. The shape of that curve is what matters. 

The curve starts steep, then levels off. That levelling is the point of diminishing returns: the moment where extra spend stops buying proportional revenue. 

The critical piece of information is where your current spend sits on that curve. The model places a marker showing your actual position. If that marker is low on the initial steep part of the curve, then you’re under-investing and additional spend would generate strong returns. If it’s up near the flat section, you’re approaching saturation and additional spend is increasingly wasteful. 

This reframes the budget conversation entirely. Instead of arguing about whether to spend more or less, you’re asking a more precise question: which channels still have steep curve remaining, and which have flattened out? The answer tells you exactly where to move budget from and where to move it to. 

In a real engagement, response curve analysis typically reveals that two or three channels have meaningful headroom for additional investment, two or three are roughly optimally placed, and one or two are clearly saturated. Reallocating from saturated to under-invested channels without increasing total spend, consistently delivers a measurable improvement in overall ROI. 

Synergy: The Multiplier Effect Most Retailers Are Leaving on the Table

The third output is the one that surprises most people when they see it for the first time: the synergy effect between channels. 

The instinct in most media planning is to evaluate channels independently: TV does X, paid search does Y, leaflets do Z. Add them up, that’s your total. But that’s not how consumer behaviour works in real life. Channels influence each other, and the combined effect of running them together is greater than the sum of their individual contributions. 

The mechanism is fairly intuitive once you think about it. A consumer sees a TV ad building awareness of a promotion. Later that week, a leaflet lands through their door confirming the deal. They search online for more information and find a paid search ad that makes the conversion easy. Each of those touchpoints individually contributed something. But the sequence was what drove the purchase. No single attribution model can see that. An MMM can. 

TV only +6.2% + Digital only +7.8% = TV + Digital combined +19.4% 

The numbers above, drawn from a real campaign analysis, show the effect clearly. TV alone generated a 6.2% sales uplift during the campaign period. Digital alone generated 7.8%. A simple addition would predict 14%. The combined campaign delivered 19.4%. That additional 5.4 percentage points is the synergy effect; it’s value that only exists because the channels were running simultaneously and reinforcing each other. 

The practical implication is significant. If you’re evaluating whether to cut a channel based on its standalone ROI, you may be undervaluing its contribution to the rest of your mix. A channel that looks marginal in isolation might be doing important work amplifying the performance of every other channel running at the same time.

Putting the Three Outputs Together

ROI analysis tells you the historical performance of each channel. Response curves tell you where each channel sits relative to its potential. Synergy analysis tells you how those channels are working together and what you’d lose by breaking up the combination. 

Used together, these three outputs support a very different kind of media planning conversation. Now it’s not just about “which channels should we cut to save money?” or “the agency says we should increase digital”, but: here is the evidence of what’s working, here is where the opportunity lies, and here are the scenarios that show exactly what happens if we change the mix. 

The three questions these outputs answer 

• ROI by channel answers: what has each channel delivered relative to its cost? 

• Response curves answer: where should we put the next pound of budget? 

• Synergy analysis answers: what happens to the whole mix if we change one part of it? 

The next article addresses one of the most common questions we hear from retailers: “This sounds right for large, single-banner businesses, but does it work for us?” Whether you’re managing multiple banners, a wholesale operation, or a network of independent stores, the answer is yes. Here’s how. 

Previous article: Leaflets: A $14,5M Retail Case Study

Next article: Multi-banner, Store-Level, Wholesome: Does MMM Work For My Business?

Continue reading the series 

Six articles taking you from the measurement problem to practical readiness — written for retail marketing leaders. 

View all articles