Marketing mix modelling
A complete MMM workflow that runs entirely in your browser.
Build the model, check whether it is any good, optimise the budget and forecast forward — without your data leaving the tab. No uploads, no infrastructure, no waiting on a data science team.
Free plan covers the full workflow on our sample datasets. No card needed.
Nothing is uploaded
The Python runtime — NumPy, pandas, SciPy — is compiled to WebAssembly and runs inside your tab. Your data is never sent to a server, because there is no server to send it to.
No account team required
The whole workflow is one sitting: load, model, diagnose, optimise. Save your work as a file you keep, and reopen it whenever you like.
Methods you can defend
Regression and conjugate Bayesian updating — fast, transparent and reproducible. Every number on screen traces back to a coefficient you can inspect.
The whole workflow
Not a demo of one step. Everything from a raw file to a defended budget recommendation.
- 1
Load and explore
Drop in a CSV. Profile every column, check correlations against the KPI, and build adstock and saturation features with a live preview of what each transform does to the series.
- 2
Specify and fit
Ordinary least squares or conjugate Bayesian regression, with priors and coefficient bounds where you need to hold a channel to what you know. A variable sweep ranks every candidate before you commit to it.
- 3
Diagnose honestly
A health scorecard grades the model across fit, residual behaviour, collinearity, coefficient sanity, contribution plausibility, transform shape and data sufficiency — and tells you what to do about each failure.
- 4
Plan and forecast
Response curves, a budget optimiser that equalises marginal returns, an efficient frontier across budget levels, and a forecast you can drive with your own plan.
Try it on our data before you trust it with yours
Four sample datasets — e-commerce, CPG, financial services and a store-level panel — run the entire workflow on the free plan. When you are convinced, Pro unlocks your own files.