All answers

The Prevailing
Answers for
Business model:
Business model, pick up to two

Should we use an open-source model (such as Meridian or Robyn) or a paid vendor?

The PrevailingUpdated Oct 29, 2026

Our answer

Pick by who will run it. Use free Meridian, Robyn or PyMC-Marketing only if a named analyst can build, refresh and test it. Otherwise pay a vendor, and check either against your own lift tests.

Confidence
Medium
Strongest proof
4 official rules
3 practitioners, 11 opinions or claims
Sources
13 sources
18 statements

New here? Take the 1-minute tour

Why this is the answer

3 reasons and our judgment

Open source is free in cash but not in time. Eightx, a fractional CFO firm, puts an open-source build at no cash but about $40K to $120K of senior analyst time, against about $1.5K to $4K a month for Recast.

In-house builds tend to fail at upkeep, not at the start. BlueAlpha and Haus, which sell mix models, argue this. BlueAlpha asks who refits the model in month seven, and Haus puts the build or buy decision around month twelve.

Paid and free models are often the same engine. AdBeacon and Michael Kaminsky of Recast say some paid tools run Meridian or Robyn underneath. Kaminsky judges any model on a holdout forecast and lift tests.

Our judgment

Choose by who will build, refresh and test the model rather than by license cost, because the estimates from Eightx and mbuzz and the maintenance failures BlueAlpha and Haus describe put the cost in people. Without an analyst already on staff, a vendor is usually cheaper than hiring one. If you build, match the engine to your analyst's language and your mix (Robyn in R, Meridian or PyMC-Marketing in Python). If your analyst could work in either, lean toward Meridian, given AdExchanger's report that Meta has stopped pushing Robyn, Robyn's latest release dating from July 2025 and Meridian's 2.1.0 release in September 2026.

Our judgment, built from Eightx, mbuzz, BlueAlpha, Haus, Impression, AdExchanger, Meta and Google: our reading of these sources, not a finding from any one of them.

What the sources say

18 statements from 13 sources
  • GoogleUndatedSupports

    Google lists certified partners, such as Adswerve, Jellyfish and Accenture, that help deploy its Meridian mix model and GeoX tests. It publishes no prices for them.

    “Choose from our certified global partners to help you deploy Meridian's advanced MMM and GeoX incrementality measurement capabilities.”

    Read it on developers.google.com

  • Meta's Robyn mix model package on CRAN is at version 3.12.1, published July 2, 2025. That is a release date, not a statement that Meta stopped work on it.

    Read it on cran.r-project.org

  • AdExchangerOct 19, 2026Caveat

    Boathouse's strategy chief says that for one client an off-the-shelf mix model credited paid media with about 75% of new customers, while a custom model on the same data said under 1%. Check models with lift tests.

    “A result below 1% deserves as much scrutiny as a result of 75%.”

    Read it on adexchanger.com

  • HausSep 30, 2026Supports

    Haus, which sells a causal mix model, says open-source models may be enough for a small expert team with a simple question, but they are not causal and come with no scientist to drive tests and adoption.

    “Open-source does not mean causal.”

    Read it on haus.io

  • GoogleSep 25, 2026Caveat

    Google's Meridian package had twelve releases in 2026, the latest being version 2.1.0 on September 25, 2026, according to its PyPI record.

    Read it on pypi.org

  • HausSep 23, 2026Supports

    A Haus scientist says a typical in-house mix model takes weeks to stand up, but clashes with tests lead teams to tune priors, and around month twelve they must fund a team or pay an outside provider.

    “And prior tuning is essentially cherry-picking.”

    Read it on haus.io

  • AdBeaconSep 22, 2026Supports

    AdBeacon, which sells a Meridian-based add-on, says any open-source mix model still needs a clean weekly data pipeline, modeling expertise and compute, and to check if a paid tool is just Meridian or Robyn underneath.

    “Confirm whether the paid tool is built on a genuinely different methodology or layered on top of Meridian or Robyn underneath.”

    Read it on adbeacon.com

  • GoogleSep 5, 2026Supports

    Google says its free Meridian mix model requires Python 3.11 or 3.12 and recommends at least one GPU.

    “We also recommend using a minimum of 1 GPU.”

    Read it on developers.google.com

  • MarSciAug 11, 2026Supports

    MarSci, which sells a managed mix model on Meridian, says build only if modeling is close to your core business and you can staff it; most mid-sized teams should buy. Either way, check it with an incrementality test.

    “Meridian and Robyn are free. The other 90 percent of the project is not.”

    Read it on mar-sci.com

  • mbuzzAug 4, 2026Supports

    mbuzz, an attribution vendor, says free Robyn and Meridian still need a modeler who costs more a year than most paid mix tools, plus 12 to 24 months of consistent weekly data. Under 12 months, run an experiment.

    “Both are free. Neither is cheap.”

    Read it on mbuzz.co

  • RecastJul 28, 2026Supports

    Recast's co-founder says free open-source mix models beat an analyst's ad hoc regression, but some paid tools just run Meridian behind attribution dashboards. First decide how you will check the model is correct.

    “is how are we going to evaluate if the model is correct or not”

    Read it on mdm.co

  • RecastJul 28, 2026Supports

    Michael Kaminsky, whose firm Recast sells mix models, says most businesses do not need one and it can mislead them; he cites studies where last-touch was about 85 percent as good as experiments in simpler mixes.

    “most businesses do not need an MMM. In fact, an MMM will be net negative for them because it will actively mislead them.”

    Read it on mdm.co

  • AdExchangerJul 14, 2026Caveat

    AdExchanger reports, from vendor and agency sources, that Meta has stopped pushing its Robyn mix model while Google's sales force drives Meridian. A rival vendor's CEO says Meridian needs customizing.

    “it’s free in the sense that a puppy might be free, not like a free beer”

    Read it on adexchanger.com

  • ImpressionJun 10, 2026Supports

    Impression, a UK agency that builds mix models, says open-source models take statistical skill and time but, done right, are more reliable than automated tools. It suggests Robyn, Meridian or PyMC by situation.

    “PyMC Marketing has no ties to any advertising platform.”

    Read it on impressiondigital.com

  • EightxJun 1, 2026Supports

    Eightx, a CFO firm citing vendor materials, says SaaS mix modeling for mid-market DTC runs roughly $30K to $80K a year, and free open-source tools cost about $40K to $120K of senior analyst time to build.

    “Recast specifically signals roughly $1.5K-$4K/month (about $18K-$48K/year).”

    Read it on eightx.co

  • BlueAlphaMay 12, 2026Supports

    BlueAlpha, a mix modeling vendor reviewing itself among 15 tools, says in-house builds fail at maintenance: build only if a named person will refit the model and run the tests. It has no data on outcomes.

    “In-house builds fail at maintenance, not construction.”

    Read it on bluealpha.ai

  • CometlyFeb 12, 2026Supports

    Cometly, which sells attribution, says a mix model needs two to three years of data and at least four channels; under $100,000 a year across fewer than four channels, focus on attribution first.

    “The practical threshold for most small businesses: if you're spending less than $100,000 annually on marketing across fewer than four channels, focus on attribution first.”

    Read it on cometly.com

  • An academic review says to gather a year of spend and sales data first, then use Robyn without in-house marketing science, Meridian with Bayesian skills, or PyMC-Marketing for experts, and check with experiments.

    Read it on thearf-org-unified-admin.s3.amazonaws.com

In the full answer

For readers who sign in
  • What to do this week3 steps
  • The numbers behind it3 figures from cited sources
  • Where each source standsA chart placing each source on the question
  • The strongest disagreementThe best case against the answer, from a cited source
  • What the research saysWhat academic research adds, kept apart from the answer
  • How sure we areWhat its confidence rests on, and what would change the answer

Read next

More in Media mix modeling