A marketing mix modeling template is a reusable starting point for one step of an MMM project: a brief that pins down the decision, a data template that structures weekly inputs, a spreadsheet or script that fits the model, a test plan that checks it, and a memo that turns the result into a recommendation. This page collects every free template in the Vexpower MMM library, from the MMM Excel template and brief template to the measurement plan and decision memo, and shows which one to use at each stage.
Most MMM templates you’ll find online are a spreadsheet model and nothing else, which skips the parts of a project that decide whether it’s useful. In my experience, building models for brands like Monzo, Booking.com and Time Out, the projects that went wrong rarely went wrong in the regression. They went wrong because nobody agreed what decision the model was for, or the data had a duplicated week nobody noticed, or a perfectly good result was presented in a way the finance team couldn’t act on. So the templates here follow the whole arc of a project, in the order you’ll need them. If you’re new to the method itself, start with what marketing mix modeling is.
Find the right MMM template for your stage
Pick the stage you’re at and the tool you work in, and the picker shows the templates that help most right now, with a direct download and a link to the guide that explains each one. Each stage ends with the question you should be able to answer before moving on, which is the closest thing I have to a quality gate for an MMM project.
You're here if: You have a budget question but no agreed decision, KPI or plan yet.
-
DOCX
MMM brief template (Word)
Write the one-sentence brief and the ten sections before anyone touches data. Guide: Measurement plan
-
MD
MMM brief template (Markdown)
The same brief for Notion, Google Docs or a repo. Guide: Measurement plan
-
XLSX
Measurement plan template (Excel)
Map each business question to a method, owner, cadence and decision threshold. Guide: Measurement plan
-
XLSX
Measurement evidence register (Excel)
Log what you already know from tests, attribution and past models. Guide: Measurement framework
-
XLSX
MMM build vs buy TCO workbook (Excel)
Decide whether to build in-house, buy software or hire a consultancy. Guide: MMM cost: build vs buy
-
XLSX
MMM vendor evaluation scorecard (Excel)
Score vendors on the same questions if you are buying. Guide: MMM software
Nothing tool-specific at this stage, so all of these work whatever you use.
Before you move on: Can you say, in one sentence, which decision the model will change and by when?
You're here if: The brief is agreed and you are collecting weekly spend, sales and controls.
-
XLSX
MMM data template (Excel)
Lay out weekly data with validation and a live audit of gaps, duplicates and zero weeks. Guide: Data requirements
-
XLSX
MMM control variable audit (Excel)
Decide which controls belong in the model and which would bias it. Guide: Confounding variables
-
ZIP
Northstar decision pack (zip)
Practise a data audit on a deliberately messy handoff first. Guide: MMM example
-
PY
Multicollinearity check (Python)
Check whether channels move together too closely to separate. Guide: Multicollinearity
-
PY
Spurious regression check (Python)
Screen trending columns before they produce fake relationships. Guide: Marketing econometrics
Nothing tool-specific at this stage, so all of these work whatever you use.
Before you move on: Does every week appear once, in one unit, with every blank explained?
You're here if: You have clean, audited weekly data and want a first model.
-
XLSX
MMM Excel template
A complete live-formula MMM: adstock, Hill, LINEST, decomposition and response curves. Guide: MMM in Excel
-
XLSX
Adstock calculator (Excel)
Try adstock shapes on your own spend before choosing one. Guide: Adstock
-
IPYNB
MMM Python tutorial notebook
A from-scratch MMM in numpy, pandas and scipy, already executed. Guide: MMM in Python
-
IPYNB
Ridge regression MMM notebook (Jupyter)
Ridge regression with a ridge trace and time-series cross-validation. Guide: Ridge regression
-
PY
Meridian quickstart on the practice data (Python)
A national Meridian model with ROI priors and diagnostics. Guide: Google Meridian
-
PY
PyMC-Marketing quickstart on the practice data (Python)
PyMC-Marketing with geometric adstock and logistic saturation. Guide: PyMC-Marketing
-
R
MMM R tutorial script
The from-scratch model in base R. Guide: MMM in R
-
R
Robyn quickstart script for the practice data (R)
Robyn inputs, run, outputs and allocator on the practice data. Guide: Robyn
Nothing tool-specific at this stage, so all of these work whatever you use.
Before you move on: Do the decomposition and ROAS look plausible to someone who knows the business?
You're here if: You have a model and need to know whether to trust it.
-
CSV
Holdout model comparison dataset (CSV)
See what overfitting looks like on a holdout before you judge your own model. Guide: Model validation
-
MD
Incrementality test brief (template)
Plan a test that checks the channel the model is least sure about. Guide: Incrementality testing
-
XLSX
Power and sample size calculator (Excel)
Work out whether that test can detect the lift you expect. Guide: Statistical power
-
XLSX
Lift test results calculator (Excel)
Read out lift, intervals and iROAS from the result. Guide: Incrementality testing
-
PY
Geo test analysis script (Python)
Analyse a geo test with before/after, DiD and fixed effects. Guide: Geo lift testing
-
PY
Synthetic control script (Python)
Build a synthetic control and run a placebo check. Guide: Synthetic control
-
PY
Calibration inputs script (Python)
Turn a lift test into calibration inputs for Robyn, PyMC-Marketing or Meridian. Guide: MMM calibration
Nothing tool-specific at this stage, so all of these work whatever you use.
Before you move on: Does the model predict weeks it has not seen, and agree with at least one experiment?
You're here if: You trust the model enough to plan with it.
-
XLSX
Breakeven ROAS & contribution margin calculator (Excel)
Turn ROAS into profit and find the breakeven point for the next dollar. Guide: Breakeven ROAS
-
XLSX
Scenario and sensitivity workbook (Excel)
Compare plans across scenarios and see which assumption moves the answer most. Guide: Scenario planning
-
XLSX
Price elasticity calculator (Excel)
Check price effects if price is part of the decision. Guide: Price elasticity
-
XLSX
Share of voice and ESOV planner (Excel)
Sense-check long-term brand budgets against share of voice. Guide: The Long and the Short of It
Nothing tool-specific at this stage, so all of these work whatever you use.
Before you move on: Does the recommendation still hold under a pessimistic assumption and a lower margin?
You're here if: You have a recommendation and need someone to act on it.
-
DOCX
MMM decision memo template (Word)
The seven-section decision memo with a review rubric. Guide: Presenting results
-
MD
MMM decision memo template (Markdown)
The same memo as Markdown. Guide: Presenting results
-
MD
Lift study readout (template)
A one-page readout for any lift study you cite as evidence. Guide: Lift studies
-
XLSX
Attribution model comparison (Excel)
Show stakeholders why last-click numbers disagree with the model. Guide: MMM vs attribution
Nothing tool-specific at this stage, so all of these work whatever you use.
Before you move on: Could a CFO repeat the recommendation, the range and what would change your mind?
Which template to use at each stage of an MMM project
The six stages below are the same ones the Vexpower MMM course follows with its invented brand, Northstar. They overlap in practice (you’ll usually plan an experiment while the first model is still being fitted), but the order matters, because a mistake at one stage is far cheaper to fix there than at the next.
Brief: agree the decision before the data
The MMM brief template is the one I’d fill in first, and it takes an hour rather than a week. It starts with a one-sentence brief (which decision, which KPI, by when) and then ten sections covering scope, channels, controls, data owners, experiments and sign-off. The measurement plan workbook takes the same thinking one level up, mapping each business question to the method that will answer it, an owner, a cadence and a decision threshold. If you already have evidence from lift tests, attribution or a previous model, the measurement evidence register is where to write it down, so the new model is judged against what you know rather than in a vacuum.
This is also the stage to decide who builds the model. The build vs buy workbook compares the five-year cost of an in-house model, software and a consultancy, and the vendor scorecard gives you the same 17 questions to ask every vendor.
Data: audit before you model
The MMM data template is a weekly data sheet with a data dictionary and a live audit that flags missing, duplicate and non-Monday weeks, blanks, zero-spend weeks, low variation per channel and how many rows you have per parameter. Pair it with the control variable audit, which asks four yes/no questions about each candidate control and suggests whether it’s a confounder to include or a mediator to leave out. If you work in Python, the VIF check and spurious regression check scripts screen for channels that move together and trending columns that produce fake relationships.
If you want to practise this stage before doing it for real, the Northstar decision pack has a deliberately messy weekly file with four planted problems, and you can preview it on the datasets page before downloading.
Model: start simple, then reach for a library
The MMM Excel template is the best place to start if you’ve never built a model, because every step is a formula you can click on: adstock, Hill saturation, LINEST, decomposition, ROI, response curves and a Solver-ready reallocation. The adstock calculator is a smaller workbook for trying carryover shapes on your own spend. In code, the Python tutorial and R tutorial build the same from-scratch model, and the ridge regression notebook shows how regularization handles correlated channels.
Once you know what a reasonable answer looks like, the library quickstarts run Robyn, Meridian and PyMC-Marketing on the same 104-week practice dataset, so you can compare what each one says about the same data. I’d always fit the simple model first; it’s much easier to spot when a library’s defaults have done something odd if you’ve seen the answer by hand.
Validation: check the model against reality
A model that fits the past well can still be wrong about which channel did the work, so validation has two halves. The first is predictive: the holdout dataset shows what overfitting looks like when you hold back the last weeks. The second is causal, which means running an experiment on the channel the model is least sure about. The incrementality test brief plans the test, the power calculator tells you whether it can detect the lift you expect, and the lift test results workbook reads out the answer with intervals and iROAS. For geo tests there are scripts for geo test analysis and synthetic control, and the calibration inputs script turns one lift result into the formats Robyn, PyMC-Marketing and Meridian accept.
Decision: turn returns into profit and stress-test the plan
The breakeven ROAS calculator converts your margins into the ROAS the next dollar needs to clear, which is the step that turns a revenue model into a profit decision. The scenario and sensitivity workbook runs the Northstar planning curves across plans, response assumptions and margins, with a tornado-style chart showing which assumption moves the answer most. If price is part of the decision, the price elasticity calculator estimates a log-log elasticity from your own price and volume data, and for long-term brand budgets the share of voice planner checks your plan against excess share of voice.
Communication: write the recommendation people will act on
The decision memo template is seven sections with a review rubric, and it’s built around the idea that the people approving a budget want the recommendation, the range of plausible outcomes and what would change your mind, not the model’s coefficients. If you’re citing a platform lift study as evidence, the lift study readout puts it on one page in a form a sceptical reader can check. And when someone asks why the model disagrees with last-click reporting, the attribution model comparison workbook shows the same journeys credited five different ways, which tends to end that conversation faster than any slide.
Every marketing mix modeling template and download on the site
The full list below is generated from every page in the library, grouped by file type, with a link to the page that explains each file. It updates itself whenever a guide adds a new download, so it’s always the complete set. All the Excel workbooks use live formulas with no macros, each script’s guide explains what data it runs on, and the example rows are invented, so replace them with your own numbers before you rely on any output.
Datasets (CSV)
Explained in What is MMM?
Explained in Model validation
Explained in Geo lift testing
Explained in MMM software
Explained in Attribution tools vs MMM
Explained in Interview questions
Explained in MMM research papers
Explained in Jin et al. (2017)
Explained in MMM datasets
Case packs
Explained in MMM example
Excel workbooks
Explained in MMM vs attribution
Explained in Measurement framework
Explained in Confounding variables
Explained in Adstock
Explained in Data requirements
Explained in Incrementality
Explained in Incrementality testing
Explained in Statistical power
Explained in Breakeven ROAS
Explained in Scenario planning
Explained in Measurement plan
Explained in Price elasticity
Explained in MMM in Excel
Explained in MMM software
Explained in MMM cost: build vs buy
Explained in Lewis & Rao (2015)
Explained in The Long and the Short of It
Python scripts
Explained in Marketing econometrics
Explained in Causal inference
Explained in Multicollinearity
Explained in Adstock
Explained in MMM calibration
Explained in Geo lift testing
Explained in Difference in differences
Explained in Synthetic control
Explained in Google Meridian
Explained in PyMC-Marketing
Explained in MMM in Python
Explained in Jin et al. (2017)
Explained in Facebook RCTs vs observational
Jupyter notebooks
Explained in Ridge regression
Explained in MMM in Python
Markdown templates
Explained in Incrementality testing
Explained in Lift studies
Explained in Presenting results
Explained in Measurement plan
Explained in Incrementality testing tools
Explained in AI and MMM
Word templates
Explained in Presenting results
Explained in Measurement plan
R scripts
Explained in Robyn
Explained in MMM in R
BIB
Explained in MMM research papers
If a term in any of these templates is unfamiliar, the MMM glossary defines it in a sentence or three and links to the guide that covers it.
