Google announced a set of measurement upgrades on September 10, 2026: a Data Strength Uplift metric that quantifies how many additional conversions your first-party data setup is recovering, a Universal Data Manager API built on IAB Tech Lab's Event and Conversions API standard, and upgrades to Meridian, Google's open-source marketing mix model, including agentic capabilities and global availability for geo-based causal experiments. The headline number Google is citing: advertisers connecting offline and app data saw a 26% average increase in incremental ROAS. All of it measures value inside Google's own walled garden (Search, YouTube, Display & Video 360), not across the other platforms your budget also runs on.
Here's the catch that matters if you run more than one platform, which is most accounts. Data Strength Uplift tells you how well your first-party data is working inside Google. It cannot tell you whether the conversion Google is crediting to a Search click was actually driven by the Meta ad the same customer saw first, because Data Strength Uplift isn't built to see Meta at all. A better-measured silo is still a silo.
- Data Strength Uplift, announced September 10, 2026, calculates additional conversions recovered by your first-party data setup inside Google's platforms specifically.
- The Universal Data Manager API is built on IAB Tech Lab's Event and Conversions API standard, meant to connect audiences across ad platforms through one setup rather than platform-by-platform integrations.
- Meridian, Google's MMM tool, gets agentic capabilities, brand-signal integration, and global availability for Meridian GeoX causal geo-experiments.
- Google's cited result: a 26% average increase in incremental ROAS for advertisers connecting offline and app data, measured within Google's reporting, not validated against a third-party or cross-platform baseline.
- None of these tools resolve the question of whose number is right when Google, Meta, and your own analytics disagree on the same conversion.
What exactly did Google announce on September 10?
Five distinct pieces, bundled under one measurement push:
Data Manager integration into GA and DV360. First-party customer data now flows automatically across Google Analytics, Display & Video 360, and Google Ads without custom engineering work on the advertiser's side. Previously this required a bespoke data pipeline per platform pairing.
Enhanced Conversions in GA and DV360. Both products now offer the enhanced conversions modeling previously more associated with Google Ads directly, with Google citing an average 11% increase in Search conversions compared to standard conversion imports.
Universal Data Manager API. Built on the IAB Tech Lab's Event and Conversions API standard, intended as a single setup to connect first-party audiences across multiple ad platforms rather than building a separate integration for each one.
Data Strength Uplift. A new measurement signal quantifying how many additional conversions your first-party data setup is recovering versus not having that data connected.
Meridian MMM upgrades. Agentic capabilities, brand-signal integration, and global availability for Meridian GeoX, which runs causal geo-experiments to validate incrementality at a market level.
Is Data Strength Uplift the same thing as incrementality?
No, and conflating the two is the mistake to avoid. Data Strength Uplift measures the value of your first-party data connection: how many more conversions Google's systems can attribute when your customer data is flowing in well versus poorly. It's a data-quality metric wearing a performance metric's name.
Incrementality measures something different: whether the ad caused the conversion at all, versus the customer converting anyway. A campaign can have excellent Data Strength (rich first-party signal, high match rates, clean enhanced conversions) and still have weak incrementality, if most of what it's "recovering" credit for was going to happen regardless. Google's own Meridian MMM upgrades are actually the more relevant tool for the incrementality question, specifically Meridian GeoX's causal geo-experiments, which test regions against each other rather than reading a platform's self-reported attribution.
The practical implication: if your team sees Data Strength Uplift climb after connecting more first-party data, that's real information about signal quality inside Google, worth having. It's not evidence the additional spend is working better, and reporting it as if it is conflates two different questions your CFO will eventually ask separately.
What does the Universal Data Manager API actually solve, and what doesn't it?
It solves an integration problem, not a measurement-agreement problem. Building on the IAB's Event and Conversions API standard means, in theory, one setup instead of custom plumbing for each destination: Google, and eventually other platforms that adopt the same standard. That's a real reduction in engineering overhead for teams currently maintaining separate server-side conversion pipelines per platform.
What it doesn't solve: whose number is authoritative once the same conversion event reaches multiple platforms. Sending the same purchase event to Google and Meta through a shared standard doesn't stop each platform's attribution model from claiming it differently, it just makes the sending easier. The reconciliation problem downstream is unchanged, which is the actual subject of cross-platform ad reporting: every platform will keep reporting its own version of the truth, cleaner data pipes or not.
How do the Meridian MMM upgrades compare to platform-reported ROAS?
Meridian is Google's open-source marketing mix model, and it's a genuinely different measurement approach from the conversion-tracking metrics above. MMM works from aggregate spend and outcome data over time rather than individual-user tracking, which makes it privacy-resilient and, in principle, blind to which platform is "supposed" to get credit.
| Platform-reported ROAS | Data Strength Uplift | Meridian MMM / GeoX | |
|---|---|---|---|
| What it measures | Attributed conversions per platform's own model | First-party data quality inside Google | Aggregate causal effect of spend on outcomes |
| Cross-platform view | No, siloed per platform | No, Google-only | Yes, if you feed it cross-platform spend data |
| Answers "did this cause the sale" | No, attributed, not causal | No, a data-quality signal | Yes, especially with GeoX geo-experiments |
| Vendor-neutral | No | No | Yes, open-source, model is inspectable |
| Setup effort | Low, native to platform | Low, layered on existing GA/DV360 | High, requires clean historical spend/outcome data across channels |
The upgrade worth noting for teams already using Meridian: agentic capabilities and brand-signal integration reduce some of the historically heavy setup lift, and global GeoX availability means causal geo-testing is no longer limited to markets where it was previously supported. That's a meaningful lowering of the barrier to a genuinely rigorous measurement method, not a replacement for it.
Should a multi-platform account switch to Google's measurement stack?
Only for the slice of the problem it actually solves: how well-instrumented your Google-side conversion tracking is, and whether your spend allocation holds up under a causal test rather than an attributed one. Both are worth having.
What it won't do is answer the same question your reporting is actually stuck on: reconciling one purchase across Google, Meta, and your own analytics into a single number you can present without three caveats. That reconciliation problem is platform-agnostic by definition. No single platform's tooling, however upgraded, is built to referee a dispute it's a party to. Google's own 26% incremental ROAS figure for this announcement is measured inside Google's systems. Treat it the way you'd treat any vendor's cited result about its own platform: useful directionally, not as a cross-platform truth.
Frequently asked questions
Related reading
- Cross-Platform Ad Reporting: Why Nothing Adds Up (And What To Trust): the cluster pillar on why every platform's number is a different version of the truth
- Incrementality Testing: The Only Measurement That Answers the Question You're Actually Asking: the causal-testing approach Meridian GeoX belongs to, outside Google's own tooling
- Your ROAS Is Lying: A Field Guide to Reading Cross-Platform Numbers: why a better-measured platform silo is still a silo


