Google Meridian Gains MMM Market Share as Meta Winds Down Robyn

Full analysis

Google is giving away the tool that decides which ads get credit for your sales — and tying its own salespeople's bonuses to how fast advertisers adopt it. Meta just quietly killed its competing giveaway. That leaves the industry's channel-measurement math increasingly authored by the biggest seller of channels. For any operator who buys, sells, or verifies media, the question isn't whether Meridian is good software. It's who gets to set the starting assumptions everyone else builds on.

What's actually being decided: not "should I use MMM" — MMM is back for good now that tracking is broken. The real decision is whether you let a media seller's model be your default measurement layer, and how much you'll pay to keep an independent check. This is a Type 1 call — hard to reverse once Meridian's assumptions are baked into your planning and your agency's white-labeled product. Forcing function: Google is embedding Meridian into Google Analytics right now, and reps are pushing onboarding this quarter.

The Market Analyst — Follow the money on who loses. The independent MMM vendors — Analytic Partners, Nielsen's mix practice, Mutinex, Ekimetrics — just watched a free competitor get wired into Google Analytics, which every marketer already has open. That eats the bottom half of the paid MMM market within two quarters, mid-market first. But there's a real contrarian trade: the conflict-of-interest story is loud enough that spooked enterprise advertisers may pay more for independence, which helps the survivors who own proprietary data or serve legally cautious clients. Meta walking away from Robyn isn't a retreat from measurement — it's a bet that in-platform signals (Advantage+) beat fighting an infrastructure war it can't win. Plainly: the referee now works for one of the teams, and the neutral referees have to prove they're worth paying for.

The Skeptic — The Enron line is great copy and the weakest part of the argument. It only holds if buyers can't see the bias. MMM in open source is inspectable — the priors, the calibration, the holdout logic are all right there in the code. Any advertiser running geo holdout tests has ground truth to check Meridian against. And the agencies screaming loudest about Google's conflicts are often selling six-figure "proprietary" models built on Meridian's own codebase. Robyn's shutdown may just mean open-source MMM got commoditized and the engineering bill stopped paying off. Sales-rep KPIs get tied to everything — that's a quota, not a conspiracy. Plainly: a rigged scale still gets caught if you weigh a known object on it, and big advertisers have known objects.

The Operator — Tuesday morning, the break is at YouTube budget allocation. Meridian's deep hooks into YouTube, Search, and Analytics data mean Google's channels tend to look better than a neutral model shows — and if your agency white-labeled Meridian without telling you, you're setting spend on an analysis Google partially wrote. The first thing that breaks at 90 days: someone reallocates seven figures toward YouTube on outputs nobody audited. The fix is boring and urgent. Audit every MMM deliverable now and ask one blunt question: is this Meridian under the hood? Then put a methodology-disclosure clause in every measurement contract. Plainly: know whose homework you're grading before you trust the grade.

The Customer / End User — The advertiser, the one whose budget is on the line, mostly didn't ask for this and often doesn't know it's happening. That's the real problem — not that Meridian is biased, but that its assumptions arrive invisibly, laundered through an agency's branded dashboard. A sophisticated advertiser with an analytics team can cross-validate. A mid-market CMO with a lean staff and a trusted agency cannot, and won't. That's exactly the buyer Meridian-in-Analytics reaches first, and exactly the one least equipped to spot the tilt. Plainly: the people most exposed to the bias are the ones with the least ability to see it.

Where they part ways:

The sharpest fight is Skeptic vs. Customer. The Skeptic says the bias is detectable, so smart buyers are fine. The Customer says detectability is worthless if the buyer doesn't know Meridian is even there — and most of the market being reached now can't run the holdout tests that would catch it. Both are right, for different advertisers. Disclosure is the hinge.

Second tension: Market Analyst vs. Skeptic on the vendors. Is the independent MMM segment getting gutted by a free competitor, or does the conflict story hand the survivors a premium pitch? The answer depends on whether advertisers act on the conflict concern or just complain about it.

What this hinges on: two things. Whether the industry forces disclosure — do buyers find out when they're running Meridian — and whether enough advertisers pay for independence to keep a real check alive. If disclosure stays murky and mid-market defaults to free, Google's priors quietly become the industry's attribution standard. If agencies get called out and independence becomes a selling point, the market splits into free-for-the-unbothered and paid-for-the-careful.

The council leans one way on the mechanism: the Strategist's read that Meridian is a distribution vehicle for Google's channel-weighting assumptions is correct, but the Skeptic is right that lock-in moves slower than the clean "Google wins measurement" story suggests. The move to verify before committing: run your own geo holdout on a live campaign and compare Meridian's YouTube credit to the ground truth. That's cheap, it's decisive, and it converts the whole conflict debate into a number you can act on.

Prediction: Before the 2027 upfront planning cycle (roughly April–June 2027), at least one of the big three agency holdcos — Omnicom, Publicis, or WPP — will publicly launch or acquire an "independent" MMM offering explicitly positioned against platform-owned tools like Meridian.

Confidence: Medium — Agencies monetize the independence pitch faster than they fix the underlying conflict.

Why: The conflict-of-interest story is now loud enough to be a sales angle, and holdcos survive by selling advisory value advertisers can't get from a platform for free. When Google embeds a free measurement default into Analytics, the agency's differentiation shifts from "we run MMM" to "we run MMM you can trust" — a pitch that requires a branded independent product to point at. Agencies have a long track record of acqui-hiring into exactly these gaps (measurement, retail media, clean rooms) the moment a client concern becomes a budget-line question. The opposite outcome — holdcos quietly white-labeling Meridian and staying silent — is less likely now that the "measurement by the seller" framing is in the trade press, because staying silent becomes a liability the moment a client asks whose model is under the hood.

Revisit by 2027-06-30: We're right if a top-three holdco launches or buys an independence-positioned MMM product before the 2027 upfronts. We're wrong if all three simply keep building on Meridian without a distinct independent offering.