Facebook (Meta) bet on Marketing Mix Modeling (MMM) to save their ads business, after Apple’s iOS14 update gave users the right to choose if they’ll accept tracking. Because it uses aggregate data, MMM doesn’t need to invade anyone’s privacy to work, so it’s future-proof to any changes in the rules, […]
How Facebook Robyn Works Read More »
At Recast, we often speak with practitioners who are trying to build their own MMM in-house. While we think there are great reasons to use a tool like Recast instead of building in-house, many people want to try it themselves before committing to working with a partner. In this post,
Disclaimer: This blog post was written by an external contributor about Google’s open source marketing mix modeling tool LightweightMMM. The approach laid out in this document is not reflective of how Recast approaches marketing mix modeling. The actual model specification can be found in the technical model specification. After Apple’s
How Google LightweightMMM Works Read More »
Disclaimer: This blog post was written by an external contributor about approaches for measuring new marketing channels with marketing mix modeling. The approach laid out in this document is not reflective of how Recast approaches modeling new channels. The actual model specification can be found in the technical model specification.
Why Marketing Mix Modeling doesn’t work for new channels (and how to fix it) Read More »
One of the core problems to solve when doing marketing mix modeling (MMM) is handling how the results from a previous period ‘carry over’ into the next period. In marketing we know that the full effect of advertising isn’t felt on the same day or week in which the campaign
Adstocks: Accounting for the Lagged Effect of Advertising Read More »
At Recast we generally recommend using spend — not impressions — to represent channel activity in a marketing mix model. This isn’t a technical limitation of the model itself, but rather reflects our beliefs about the critical limitations of “impressions” metrics that modelers should be aware of. This recommendation is
The case against using impressions in MMM Read More »
Scenario planning or demand forecasting decides everything an enterprise does. Right from quantities of raw material sourced to the inventory that needs to be stocked, through to shipping, packaging, hiring, and all the way down to purchasing office supplies — everything is dependent on accurate forecasting and scenario planning. Forecasting
Marketing Mix Modeling for Scenario Planning Read More »
You cannot double spend and expect sales to double. Advertising on auction-based platforms like Meta or Google exhibits diminishing returns, meaning as you increase spend, efficiency suffers. When you continue to increase your advertising spend you see increased customer acquisition cost (CAC) and lower return on investment (ROI). Essentially, there’s
Diminishing Returns: Accounting for Channel Saturation Read More »
Disclaimer: This blog post was written by an external contributor about detecting and addressing heteroskedasticity in marketing mix models. The approach laid out in this document is not reflective of how Recast approaches heteroskedasticity and model validation. The actual model specification can be found in the technical model specification. Heteroskedasticity
How to fix heteroskedasticity in your Marketing Mix Model Read More »
Multicollinearity is one of the hardest problems in marketing analytics. If you took a statistics course in undergrad, you might have covered it; it describes a phenomenon where two or more variables in a model are correlated. Let’s bring it to marketing to make it more tangible. Imagine we’re trying
How D2C Brands Can Deal with Multicollinearity in their Marketing Mix Modeling Read More »
