Marketing mix modelling

Choose a measurement question, audit your data and learn how an MMM turns channel activity into a budget decision.

Start with a decision you could actually change: next quarter’s channel budget, whether to expand a market, or how much evidence you need before scaling a campaign. A model is useful when it narrows that decision, even if it cannot give every channel a precise return.

Begin with What is MMM? and Data requirements. Then connect the analysis to the MMM course exercises, where you practise making a decision from model outputs.

Before acting on a recommendation, connect this topic to budget decisions: the result needs both credible measurement and a commercially worthwhile action.

A practical route through the topic

1 Define the decision
2 Audit weekly data
3 Fit and challenge
4 Test a small change

Suppose sales rise during a promotion while search spend also rises. Giving search all the credit would confuse demand, price and advertising. A useful model includes those competing explanations and checks whether its channel estimates survive different reasonable specifications.

Try it before reading further

Change the inputs below and explain what changed. These teaching tools illustrate assumptions; they do not estimate your business’s return from the information on this page.

Decision you need to make

Monthly paid media budget

Channels involved

When you need the answer

Data you already have 2+ years of weekly spend and sales by channel Sales by region or market Reliable click and conversion tracking A brand tracker or regular survey Results from past lift or geo tests

What to check before using the result

A good fit is not proof of causality. If two channels always move together, more rows of the same pattern may not separate their effects. Record what the data cannot identify before presenting a recommendation.

One outcome, currency, geography and weekly calendar agreed with finance
Spend, sales, promotions and material changes reconciled to source systems
A future period held back before tuning the model
A bounded decision and a plan to check the result

Your first reading path

For an overview, visit marketing mix modelling training. If you use AI to assist with analysis, the prompt engineering learning page explains the complementary skill.

Step 1 What is MMM? Marketing mix modeling (MMM) uses regression on weekly sales and spend to estimate what each channel adds. How it works, what you get, its limits, how to start. Step 2 Data requirements Marketing mix modeling data: the outcome, media and control variables you need, weekly vs daily, how much history is enough and why variation matters. Step 3 MMM example A worked marketing mix modeling example: write the brief, audit messy weekly data, challenge the model review, compute the economics, write the decision memo. Step 4 Model validation How to validate a marketing mix model: in-sample vs holdout error, rolling-origin cross-validation, MAPE, NRMSE, DECOMP.RSSD, R², stability and experiments.

Foundations from Mike’s learning briefs

Segmentation Analysis Splitting data can reveal variation but also leaves smaller samples, so choose segments around a decision. Marketing Attribution Models Attribution assigns observed credit, while experiments and aggregate models answer different measurement questions. Data Visualization A useful model chart starts with trustworthy data and makes the decision, units and uncertainty visible. Data Cleaning A weekly model needs reconciled inputs; filling gaps can otherwise create evidence that was never observed. Interpolate Data Interpolation changes the information available to a model; distinguish a missing observation from genuine zero activity. Search Trends Search interest may reflect underlying demand or be changed by advertising, so its role as a control needs care.

Turn understanding into practice

Your next step: make a better marketing decision.

Work through realistic business problems with guided lessons, practical templates and simulations.

Choose your learning plan Explore the MMM course See the curriculum before choosing a plan.

Explore every guide in this topic

Apply it to your marketing

B2B MMM: model pipeline without confusing sales-cycle lag Choose B2B outcomes, align media with pipeline maturity and avoid treating a handful of large deals as precise evidence of channel returns. Ecommerce MMM: separate promotions, media and returns Build an ecommerce MMM around net sales, promotional changes and returns timing, with a worked example of the revenue definition that changes the decision. Podcast MMM: align ad delivery, lag and promo-code evidence Build a podcast measurement plan that separates delivery estimates, promo-code attribution and incremental response before using MMM to change spend. Retail media MMM: measure halo without double-counting sales Define the scope of retail media returns across advertised products, other products and retailers, then reconcile overlapping outcomes before budget allocation. Retail MMM: account for stockouts and distribution changes Avoid crediting media for distribution growth or penalising it for unavailable products. Build a retail MMM with availability checks and bounded decisions. Travel MMM: align booking dates, stays and cancellations Choose the right outcome calendar for travel MMM and distinguish advertising carryover from booking windows, cancellation timing and capacity constraints. Industry marketing: build a useful vertical strategy Choose a target industry, adapt the offer to its buying process and measure qualified opportunities instead of broad engagement. Advertising claims: build an evidence and disclosure workflow A practical workflow for checking advertising claims, disclosures and campaign evidence, with links to UK CAP and US FTC guidance. Multi-channel campaigns: coordinate delivery and measure the total Plan a multi-channel campaign with a shared brief, consistent measurement and a clear distinction between attribution and incremental impact.

Frequently asked questions

Where should I start learning marketing mix modelling?

Start with the business decision and a worked example, then audit the dataset before choosing software. Learn validation and experimental calibration before treating channel coefficients as budget instructions.

Does an MMM replace campaign reporting?

No. Campaign reports help monitor delivery and diagnose changes. MMM estimates aggregate relationships over time, while experiments provide additional causal evidence. Use the methods together.

Connect the ideas

Incrementality and experiments Design holdouts and geo tests, understand statistical power and use experiments to challenge attribution and calibrate MMM. How marketing mix models work Explore adstock, saturation, baselines and confounding with visual explanations and interactive model mechanics. Marketing budget decisions Compare marginal returns, contribution margins and bounded scenarios before moving marketing budget. MMM tools, code and platforms Choose between R, Python, spreadsheets and managed MMM platforms, with tutorials, datasets and validation checkpoints.