MMM Analyser
Marketing mix modelling estimates how much each channel’s spend adds to sales, using the history of both. It captures two effects a simple regression misses: advertising keeps working after the week it runs, and each extra unit of spend adds less than the one before.
The model covers promotion spend and price. Product, place, people, process, and physical evidence are outside it, and are planned in the 7Ps Marketing Mix Planner.
Everything above runs in your browser and saves to this device only. Spend and sales data never leave the browser. It opens on 2 years of weekly hotel booking revenue, generated from known settings, so you can see how closely the model recovers them.
How it works
For each channel, spend is carried forward by an adstock rate, then passed through a saturation curve. Sales are modelled as a base, plus trend, seasonality, and price, plus the channel effects.
Saturation h(A) = A ÷ (A + K), with K a multiple of the channel's mean adstock
Salest = b0 + trend + seasonality + γ × pricet + Σ βc × hc(Ac,t)
ROI = contribution ÷ spend. Marginal ROI = extra sales from one more unit of spend at the current level
For fixed θ and K the model is linear and is fitted by ordinary least squares. θ and K are chosen per channel by a search that maximises R². The tool warns when the data are too short for the number of coefficients, when a channel’s effect is negative, and when two channels’ spend moves together so closely that their effects cannot be separated. Treat results as a starting point for testing budget shifts.
Frequently Asked Questions
What is marketing mix modelling?
Marketing mix modelling (MMM) is a statistical method that estimates how much each marketing channel's spend adds to sales, using the history of both. It captures carryover, where advertising keeps working after it runs, and diminishing returns.
What is adstock in marketing mix modelling?
Adstock is the carried-over effect of past advertising. With a carryover rate of 0.6, 60% of last week's effect carries into this week, then 60% of that into the next week.
What is marginal ROI in marketing mix modelling?
Marginal ROI is the extra sales from one more unit of spend at the current level. A channel with a high average ROI can have a low marginal ROI near saturation, which is the point to move budget elsewhere.
How much data does marketing mix modelling need?
Marketing mix modelling needs at least 10 periods for every coefficient. With four channels, trend, seasonality, and price, that is about 90 periods, or roughly 2 years of weekly data.
Is this MMM Analyser free?
Yes. The MMM Analyser is free and opens straight away in any modern browser, including on a phone. Spend and sales data are processed in your browser and stay on your device.