Google published open beta documentation for the Meridian Scenario Planner on 6 August 2026, a day after posting a video tutorial on 5 August 2026, according to PPC Land. The Scenario Planner is part of Meridian, Google's marketing mix modelling tool launched on 19 February 2026, and it automates the creation of reports that analyse how effective each marketing channel is and what it contributes to sales or other KPIs, delivered through an interactive dashboard rather than a static report. This is documentation for an open beta, and Google has stated no general availability date.
What Google published on 6 August 2026
PPC Land reported on 6 August 2026 that Google had released open beta documentation for the Meridian Scenario Planner, following a video tutorial posted on 5 August 2026. The reporting is on PPC Land. Meridian itself launched on 19 February 2026 as Google's marketing mix modelling tool, so the Scenario Planner is a layer on top of a model that already exists rather than a new measurement product.
The stated purpose of the Scenario Planner is budget optimisation delivered interactively. Instead of a report that tells you what happened, the dashboard is meant to answer what happens to modelled outcomes if spend moves between channels. The documentation includes a $150 million demo scenario for working through that.
What the beta documentation covers
The table below lists the areas the beta documentation addresses, as described in PPC Land's 6 August 2026 report.
| Area | What the beta documentation covers |
|---|---|
| Model fit diagnostics | R-squared, MAPE and wMAPE |
| Channel contribution | Contribution analysis with baseline separation |
| Return metrics | Three metrics: ROI, effectiveness and marginal ROI (mROI) |
| Budget optimisation | Fixed and flexible optimisation, with target total ROI and target channel mROI |
| Spend constraints | Published formulas based on minimum and maximum spend shift ratios |
Two further areas sit outside the table. The documentation gives guidance on interpreting response curves, and it sets out data security protocols that distinguish temporary dashboard links from persisted ones, with saved dashboards running on the owner's own data access rights.
Publishing the spend constraint formulas is the notable part
Google published the spend constraint formulas themselves, based on minimum and maximum spend shift ratios, rather than only describing the outcome. That is the difference between a tool that hands you a recommended reallocation and a tool whose reallocation logic you can check. An optimiser that will not show its constraints has to be trusted; one that publishes them can be argued with, which is a better position for anyone signing off a budget move.
The three return metrics matter for the same reason. ROI, effectiveness and marginal ROI are listed as separate measures in the documentation, and the distinction is not cosmetic when a budget decision is on the table. Blended return describes money already spent. Marginal return describes the next unit of budget, which is the only one you can still decide about. A channel with strong average ROI and weak marginal ROI is a channel that has run out of room, and averages hide that.
Marketing mix modelling has become the measurement approach advertisers fall back on as click-level attribution degrades under consent limits and cross-device gaps. Advertisers still need clean conversion data feeding the model, so a properly configured analytics foundation is a prerequisite rather than an alternative. A model fitted to bad inputs produces confident and wrong reallocation advice.
What this means for Thai marketers
Google has stated no Thailand-specific availability for the Scenario Planner beta, and PPC Land's report contains nothing about Thai access. Treat it as documentation you can read now, not a feature you can assume is switched on for your account.
The fit question matters more than the availability question here. Marketing mix modelling needs a long spend history across several channels running at once, which points at Thai advertisers spending meaningfully across Google, Meta, LINE and TikTok in parallel. A single-channel account has nothing for the model to compare, and a three-month history gives it too little to separate a channel effect from a seasonal one. Advertisers already running multi-campaign Google Ads programmes alongside other platforms are the ones for whom a scenario planner has something to plan.
Questions advertisers are asking
Is the Meridian Scenario Planner generally available?
No. This is open beta documentation, and Google has stated no general availability date. Reading the documentation is not the same as having the feature.
Is it available in Thailand?
Google has stated no Thailand-specific availability, and the source does not say. Nothing in the 6 August 2026 documentation release addresses country-level rollout.
What is the difference between ROI and marginal ROI here?
The documentation lists ROI, effectiveness and marginal ROI (mROI) as three distinct return metrics, and it also supports optimisation against a target total ROI or a target channel mROI. As a matter of how budget decisions work rather than something the source defines, blended return summarises spend that has already happened while marginal return speaks to the next unit of budget, so mROI is the number that argues for or against moving money.
Do I need Meridian to use the Scenario Planner?
The Scenario Planner is documented as part of Meridian, Google's marketing mix modelling tool launched on 19 February 2026. Beyond that relationship, the source does not spell out access requirements.
Is my data safe if I share a dashboard?
The documentation includes data security protocols that distinguish temporary dashboard links from persisted ones, and saved dashboards run on the owner's own data access rights. That means a shared saved dashboard inherits the permissions of whoever saved it, so who owns a dashboard is a security decision as much as an admin one.
If you are weighing whether marketing mix modelling fits your account, the honest first question is whether your conversion data is clean enough to model. Relevant Audience can look at your measurement setup and tell you what a model would and would not be able to see.
