What is a channel hierarchy in MMM?

Marketing spend arrives from twenty places in twenty shapes: an ad platform's API, an agency's media plan, a broadcaster's spot report, an affiliate network's invoice. Before any of it can be modeled it has to land in one structure. The hierarchy is that structure: markets at the top, products under them, segments where the business uses them, and then the media hierarchy, from channel type to platform to campaign to the individual spot. Search, then Google, then Performance Max. TV, then the broadcaster, then the spot.

The model works on activity types, and the hierarchy is where those are defined. Which campaigns count as one channel, whether Meta is split into awareness and conversion, whether two small display networks are grouped into one, all of that is a hierarchy decision, and it is made by how much spend each branch carries. A channel with heavy spend can be split finer; a channel with little spend is grouped with its neighbours, because the model cannot say anything reliable about it on its own.

Because the hierarchy is the same for every source, offline and digital are compared on the same terms. A TV spot and a search campaign sit in the same tree, with spend in the same currency at the same daily grain.

Odins structures every source into the customer's hierarchy during onboarding: what the business already reports on, markets and product lines first, with the media tree under it, built so that everything lands in the right place down to each spot. Sales data is structured next to the spend in the same tree, so a market's revenue sits against that market's media. The structured dataset can be delivered back to the customer's own warehouse in the same shape, and the raw platform tables with it.

• Structured marketing data: Every source, one dataset, your hierarchy, your warehouse. The foundation the model sits on, and a product on its own.
• Offline media spend: TV, radio, out-of-home, print, influencer. No API, so it needs a pipeline, down to the individual spot.
• Marketing mix modeling (MMM): The statistical method behind everything else on this list: spend in, sales out, every channel on equal terms.
• Channel contribution: What each channel added to sales above the baseline over a period. The decomposition behind every ROAS figure.

Go one level deeper: Data integration. The connectors. Or go back to all terms in the glossary.