On January 29, Google announced the wide launch of its Meridian open-source marketing mix modeling (MMM) solution, available through GitHub. Cue a flurry of LinkedIn takes speculating about the future of the marketing measurement and optimization (MMO) solutions market.
By providing an open-source Bayesian option, Google brings the industry-standard methodology to the masses (Meta’s Robyn uses a different regression process). But will Google’s free solution replace what many brands invest thousands of dollars in? No. Though there are many good things about this release, and it potentially broadens the market for MMM, it’s not a “one size fits all” solution.
Since its initial Meridian announcement in 2024, Google has clarified the availability of additional capabilities, including budget optimization, scenario planning, the ability to customize parameters according to prior results, and the inclusion of non-media and control variables. Perhaps most importantly, Meridian is distinct in its native inclusion of Google search query volume data and reach/frequency data for YouTube. Along with the code release, Google also launched a partner program with major agency and MMO partners throughout the world to help brands with implementation.
Meridian Is Best For Companies With Strong Data Science And Storytelling Chops
To DIY MMM, you need strong in-house data engineering, data science, and visualization resources. Digital agencies adept at campaign measurement but lacking an MMM solution will benefit from using Meridian to create solutions on behalf of clients. But for brands, the calculus is a little more complicated. Consider going it alone with an open-source solution such as Meridian if:
- You spend a lot of your marketing budget with Google. The ability to incorporate Google Query Volume and YouTube reach/frequency data increases in value as you spend more on these channels. If you are a digital brand that spends a majority or plurality on these channels, Meridian has significant value.
- Your in-house data team cooks. You have a team of data scientists who can clean, transform, load, and migrate sales, marketing, economic, and media data quickly and consistently. Accurate measurement requires extensive data expertise and resources. On top of that, a culture of trust between marketing and other departments (e.g., finance, sales, or operations) is necessary; you will need skilled data storytellers to translate results into compelling messaging.
- You have a history of maintaining geo-level marketing, sales, and competitor data. MMM is a wonderful tool for finding the causal relationships between variables, but its impact is dulled without enough data. MMM becomes more accurate and more useful with more granular data, particularly at the geo level. In addition to knowing what you were spending on media in Phoenix during the second week of August 2023, do you also have data related to non-media programs (such as discounts, bulk promotions, new packaging) running in that market concurrently? What about promotions that your competitors were running in the same market?
- You have experience with incrementality testing and ROI calculation. MMM results sometimes contradict what is found when measuring campaign-level ROI and return on ad spend. In general, this is because campaign-level measurement can’t account for what the result would have been if no campaign had been run. It’s for this reason that Forrester recommends a layered measurement framework, with campaign-level measurement and incrementality testing feeding into MMM. Meridian enables marketers to customize parameters to reflect results of incrementality tests and campaign ROI measures, but this is only useful if your brand has experience with these tests.
Brands Without A Measurement-Driven Marketing Culture Will Struggle To Leverage Meridian
In addition to the criteria above, a major challenge faced by many marketing departments is earning trust from other parts of the organization that their efforts are driving real business results. While Meridian’s code is open source and the underlying calculations are all laid out, the materials are not executive-friendly. Though this release answers some of the questions we raised when Meridian was announced in 2024, the education aspect remains unaddressed. If you are looking for someone to communicate the benefits, strengths, and, yes, weaknesses of MMM, Google’s materials will not help you. In addition to customized modeling and consulting, many MMO providers provide guidance for communicating findings and gaining cross-org trust.
Later this year, I’ll be publishing reports on data requirements for marketing measurement, best practices for MMM, and building strong measurement teams. If you want to discuss your MMM or other marketing measurement initiatives, schedule a guidance session here.
