Marketing mix modeling had a strange week in the trade press. On July 14, AdExchanger published "Google's Meridian And Meta's Robyn: A Gift To Measurement Or Trojan Horses?", followed three days later by a companion piece, "Picking An MMM." Both landed on the same uncomfortable question: now that Google and Meta each give away a marketing mix model for free, should brands actually build their measurement stack on top of one?

It's not an abstract question. According to AdExchanger's reporting, Google has tied internal sales KPIs to Meridian adoption — account teams are incentivized to get brands onto Google's own MMM. Meanwhile, Meta is reportedly winding Robyn down behind the scenes, even though no formal deprecation notice has been issued. One platform is pushing its model harder than ever; the other is quietly letting its walk away. Neither behavior looks like stewardship of a neutral measurement standard — it looks like two companies managing a product line.

That's worth sitting with, because MMM's entire value proposition is that it's supposed to be the one measurement method immune to platform self-interest. Unlike pixel-based attribution, MMM doesn't ask a channel to grade its own homework. It's aggregate, causal, and channel-agnostic by design. When the model itself is authored and maintained by one of the channels being measured, that promise gets shakier — even before you consider what the model can and can't see.

Why platform-owned MMM has a structural conflict of interest

AdExchanger's sourcing is blunt about this: "platform MMM is the same as platform anything. It's there to prove the platform succeeded, as much as that your marketing worked." Google's Meridian is genuinely strong at tying together Search, YouTube, TV bought through Google, and Google Ads — which is also, not coincidentally, a good way to demonstrate that the "Googleverse" drives outsized value. The article also notes that several vendors have quietly built their "proprietary" measurement offerings on top of Meridian's open-source code without clients realizing the model underneath isn't independent.

None of this means Meridian or Robyn are badly engineered — both are legitimate, well-documented Bayesian MMM frameworks, and Google added a no-code Scenario Planner interface to Meridian back in February specifically to make budget-scenario testing accessible to non-technical marketers. The issue isn't code quality. It's incentives: a model built by a company that also sells performance media has a built-in reason to find that performance media works.

The offline blind spot no platform tool can fix

There's a second, quieter problem that gets less attention in the "gift or Trojan horse" framing: platform MMM tools are built around the data the platform itself can see. That's digital by construction. Marketers have grumbled for years that Google and Meta's own reporting takes credit for conversions that actually originated from a TV spot, an out-of-home campaign, or a print insert — channels the platform has no visibility into and therefore can't model well, if at all.

That's the gap that matters most for any brand still spending meaningfully offline. A model optimized to explain Search and YouTube performance isn't optimized to capture the halo effect of a national TV flight or the local lift from OOH — and if it can't see those channels cleanly, it can't allocate budget across them accurately either. You end up with a measurement stack that's precise about the channels the vendor sells and vague about everything else, which is exactly backwards from what a strategic budget-allocation tool should be.

"Platform MMM is the same as platform anything. It's there to prove the platform succeeded, as much as that your marketing worked."

Signals that the market agrees: money and mandates

This isn't just a trade-press narrative. EMARKETER and TransUnion's most recent survey data on measurement found that 46.9% of US brand and agency marketers plan to invest in MMM over the next year, and 27.6% now call it the single most reliable measurement methodology they use — ahead of multi-touch attribution (19.4%) and unified/holistic measurement (18.9%). The same survey found 67.4% of marketers say improving incremental ROI has become more urgent given the current economy. Marketers are voting for MMM with their budgets precisely because they want the platform-agnostic view — which makes it worth asking whether a platform-authored model actually delivers that.

Capital is moving too. In June, Brazilian measurement platform Uncover raised a $16 million Series A led by Cloud9 Capital specifically to expand its always-on MMM — built to unify online and offline channels in one model — into the US market, on the strength of already helping advertisers like Unilever and Burger King optimize more than $6 billion in combined online and offline media spend. Investors are backing independent, cross-channel measurement at the same moment the two largest platforms are jockeying over their own free tools.

What "vendor-neutral" should actually mean in an MMM

Put together, the picking-an-MMM moment the industry is having right now points to three things brands should actually demand from a model, rather than defaulting to whichever platform's tool their rep is pushing hardest:

  • No channel gets home-field advantage. The model should weight and calibrate TV, OOH, radio, print, and digital using the same statistical rigor — not treat offline as an afterthought bolted onto a digital-first framework.
  • The incentives are aligned with the brand, not a media seller. A vendor with no stake in which channel wins the budget has no reason to shade the results.
  • It's always-on, not a one-off project. Static, quarterly MMM reports can't keep pace with budget decisions being made weekly; an always-on model recalibrates as new spend and outcome data arrives.

The takeaway

Google and Meta making MMM tooling more accessible is, on balance, good for the industry — it's part of why MMM adoption is accelerating heading into 2026. But accessibility and neutrality are different things, and this month's reporting is a useful reminder that a model built by a media seller will always have a seller's-eye view of what worked. For brands that genuinely spend across offline and online, the more durable answer is a causal, always-on MMM built by a vendor with no channel to protect.

That's the gap Animo was built to close. Animo measures TV, OOH, radio, print, and digital in a single causal, always-on model — with no platform stake in the outcome — so budget decisions reflect what actually drove results, not what a media seller's own tool is incentivized to show. If you're reassessing your measurement stack in light of this month's Meridian and Robyn coverage, get in touch with Animo to see what a genuinely unified view of your offline and online spend looks like.

Sources