MMM, without the blackbox
The questions we hear before, during and after a measurement project. If yours is not here, ask it in a demo.
The project
Pricing, timelines, scope
How much does a mediaROI MMM cost?
mediaROI pricing depends on your scope: the number of brands, countries and channels modelled, how often the models are updated and the level of support. Every proposal is built to measure.Contact us for a quote →
How long before the first results?
It mainly depends on the availability of your data. To speed this step up, we have built 40+ native API connectors to your media, analytics and CRM platforms. Once collection is validated, modelling and delivery take a few weeks, versus several months for a traditional MMM.See all connectors →
Does the MMM replace my attribution tool?
No, it makes it reliable. The MMM is your source of truth (online, offline and external factors) and recalibrates your attribution, which on its own overvalues the last click. They work together, not against each other.See the five steps →
Who works with us during the project?
A marketing scientist works with you on the setup, the reading of the results and the arbitrations that follow. The platform runs continuously, and the people who know your business stay in the loop.See the platform →
Do you need a data team to use mediaROI?
No. The platform is built for marketing teams, supported by a mediaROI Marketing Scientist and Data Scientist. A data team can go further with settings, training and API or MCP access.
The data
What you need to provide
What data do you need?
Your media spend and performance, the objective you want to model — revenue, new customer acquisition, leads — and what drives your sales without being media: promotions, pricing, seasonality, store openings. The connectors pull media and analytics; the rest often comes in a file, and that is expected.See the connectors →
What if a source has no connector?
If it comes in a file, it goes into the model: offline media plans, in-store data, partner exports. New API connectors are added as clients need them.
Who checks that the data is right?
Every source passes through a data checklist and a validation step. A gap, a duplicate or a currency mismatch is caught there, not three months later in a result nobody can explain. Our team standardises the naming and reconciles what two platforms count differently.
The model
Cadence, reliability, objectives
Which model powers your MMM?
On an open-source Bayesian model. Its code is public and the settings applied to your data are visible in the platform.
How often is the model updated?
Monthly, quarterly, twice a year or yearly: you choose the cadence, and that is also what determines the subscription price.
How do I know the model is reliable?
Every model ships with its quality indicators: R², MAPE, confidence intervals and the priors it was given. Your data team can inspect how a number was built and challenge it. A ROI figure is only worth something when you can explain how it was obtained.See the platform →
Can I model more than one objective?
Yes. One model per objective, and the allocations compare side by side instead of averaging out into one compromise.See the simulator →
What are my incrementality tests used for?
Your geo and conversion lift tests feed the model and correct it against reality. That is the third method in the loop: MMM measures everything, attribution follows the day to day, incrementality tests reality.
AI
mediaAI and the MCP server
What can you ask mediaAI?
Questions about your own results: contributions, ROAS, saturation, model quality. mediaAI does not invent a number it cannot trace, and it points to the view the answer came from, so anyone can check it.mediaAI →
Are my results exposed to other accounts?
No. The MCP server exposes your own measurement results, under your own access rules. Nothing is shared across accounts.MCP and AI agents →
Discover the platform
30 minutes to walk through a complete model and ask your questions