Google is moving its advertising measurement stack further toward a model built around first-party data, cross-platform activation and causal proof.

On September 10, 2026, Google announced a new set of measurement updates spanning Data Manager, enhanced conversions, the Data Manager API, a new Data Strength Uplift metric and upgrades to Meridian, its open-source marketing mix modeling framework.

The company’s message is clear: AI-driven campaign systems need better first-party signals, and advertisers need more than platform attribution to understand whether those systems are creating incremental business value.

For operators, the update matters because measurement is becoming less of a reporting layer and more of an input layer for campaign optimization.

Google is consolidating first-party data activation

Google said Data Manager is being integrated directly into Google Analytics and Display & Video 360.

That expands the role of Data Manager from a Google Ads-specific data connection layer toward a broader activation layer across the company’s advertising and analytics products.

Google also said enhanced conversions are being expanded into Google Analytics and DV360.

Enhanced conversions use hashed first-party customer data to improve conversion matching.

In practical terms, the direction is toward one operating model:

  1. collect first-party customer events;
  2. normalize them;
  3. route them through a central connection layer;
  4. use them for measurement and campaign optimization.

That can reduce the number of one-off pipelines marketers maintain across advertising products.

It also increases the importance of data governance.

If the source data is weak, centralizing it does not make it better.

Google is pushing a universal Data Manager API

Google said the Data Manager API is now universal.

The company describes it as a secure setup for connecting, managing and activating audience and measurement data across major advertising platforms.

The API is based on the IAB Tech Lab’s Event and Conversions API standard.

For larger advertisers, that is strategically significant.

Conversion architecture is moving away from a collection of platform-specific uploads toward reusable first-party event infrastructure.

A mature setup can eventually support:

  • browser events;
  • server-confirmed purchases;
  • CRM stages;
  • offline conversions;
  • app events;
  • audience activation.

The operational benefit is not only better measurement.

It is fewer inconsistent definitions.

If one API pipeline sends “qualified lead” to several platforms, the organization has a better chance of optimizing against the same business event everywhere.

The new Data Strength Uplift metric attempts to quantify signal value

Google is introducing a Data Strength Uplift metric in Google Ads.

The company says the metric estimates additional conversions recovered through a stronger first-party data setup.

Google also published internal performance figures alongside the launch.

It says advertisers connecting offline and app data to Data Manager see an average 26% increase in incremental ROAS.

Google also says enhanced conversions produce an average 11% increase in Search conversions compared with standard conversion imports.

And it reports an average 14% conversion uplift for advertisers using Google tag gateway, with more than 20% uplift for Demand Gen campaigns in one cited dataset.

These figures should be treated as Google-reported benchmarks, not universal expectations.

The company cites internal global datasets, and advertisers should not assume the same lift will occur in their own accounts.

The more useful takeaway is that Google wants marketers to measure the performance contribution of data infrastructure itself.

That is a notable shift.

Historically, tag quality was treated as technical hygiene.

Google is increasingly positioning it as a growth input.

Meridian is becoming more operational

Google also announced upgrades to Meridian, its open-source marketing mix model.

The company says it is adding agentic capabilities to help:

  • audit data quality;
  • resolve modeling errors;
  • guide model construction.

It is also making backend changes intended to speed analysis.

One important addition is support for brand signals such as branded Google Query Volume directly inside Meridian models.

That allows marketers to model a longer chain:

upper-funnel media → brand demand → future commercial outcomes.

This is useful because brand advertising is often evaluated too narrowly.

A campaign can create demand that appears later through:

  • branded search;
  • direct traffic;
  • retail;
  • sales.

MMM attempts to capture some of those effects.

GeoX is now generally available

Google said Meridian GeoX is now generally available globally.

GeoX is Google’s open-source library for running geographic experiments.

Geo experiments compare treatment and control regions to estimate incremental effect.

This can be useful when user-level randomization is difficult.

Examples include:

  • television;
  • YouTube;
  • paid social;
  • search budget changes;
  • regional promotions.

The important distinction is between attribution and incrementality.

Attribution asks which touchpoints receive credit.

Incrementality asks what happened because the marketing activity existed.

Google is increasingly combining these two layers.

The new measurement stack has three parts

First-party data sources feeding a central measurement layer and branching into performance, modeling and geographic experiment views.
Google’s new measurement framework connects data foundation, observational signals and causal proof.

Google explicitly frames its strategy around:

  1. strong data foundation;
  2. multiple signals;
  3. causal proof.

That maps well to how modern marketing teams should think about measurement.

Data foundation

Can the business reliably capture:

  • purchase;
  • qualified lead;
  • activation;
  • revenue;
  • customer identity?

Observational measurement

Can the business understand:

  • channel performance;
  • attribution;
  • funnels;
  • cohorts;
  • contribution?

Causal measurement

Can the business test:

  • lift;
  • incrementality;
  • geographic effects;
  • true media contribution?

No single tool answers all three.

The value of the new Google stack is that these layers are becoming more connected.

What operators should review now

1. Audit first-party conversion events

Ask:

  • Which events currently feed bidding?
  • Are they business outcomes or proxy actions?
  • Are online and offline outcomes connected?
  • Are values accurate?

2. Review Data Manager

If data is uploaded through separate integrations, determine whether Data Manager can simplify the architecture.

3. Check enhanced conversions

Confirm:

  • eligibility;
  • implementation;
  • consent;
  • data quality;
  • diagnostics.

4. Separate attribution from causal proof

Do not interpret a better attribution report as proof of incrementality.

Use:

  • holdouts;
  • lift studies;
  • geo experiments;
  • MMM.

5. Document source-of-truth systems

Revenue may come from billing. Opportunity stage may come from CRM. Behavior may come from analytics.

The measurement architecture should define which system owns each business state.

The strategic implication

AI-driven advertising makes measurement quality more consequential.

When campaign systems automate:

  • targeting;
  • bidding;
  • creative selection;
  • budget allocation;

the business increasingly controls performance through the quality of the inputs and constraints it provides.

That means measurement infrastructure is becoming part of media buying.

A marketer who improves the conversion signal is not merely fixing reporting.

They may be changing what the bidding model learns.

Google’s September update formalizes that direction: first-party data, modeling and causal measurement are being brought into one operating framework.

The next competitive advantage may not be another campaign optimization trick.

It may be having a cleaner, more economically meaningful signal than the advertiser competing in the same auction.

A practical 30-day implementation sequence

Operators do not need to rebuild the entire measurement stack at once.

A focused sequence can start with the conversion events that influence the most spend.

Week 1: inventory the current signal

List every primary conversion used by:

  • Google Ads;
  • GA4;
  • DV360;
  • CRM reporting.

For each one, document:

  • source;
  • owner;
  • value;
  • latency;
  • whether the event is online or offline.

Week 2: reconcile the business outcome

Compare platform conversions with the authoritative commercial system.

For ecommerce: orders and settled revenue.

For B2B: qualified opportunities and closed-won revenue.

For subscriptions: activated paid accounts.

Week 3: improve first-party connectivity

Review whether Data Manager, enhanced conversions or API-based routing can close measurable gaps.

Do not add a data connection unless the team can explain which problem it solves.

Week 4: choose one causal test

Select one meaningful allocation question.

Examples:

  • does upper-funnel video create incremental demand?
  • does a regional media increase produce incremental sales?
  • does paid social add customers beyond organic demand?

Use the simplest credible experimental design.

This turns Google's new measurement stack from a product announcement into an operating change.

The most important outcome is not enabling every new feature.

It is improving the chain between customer behavior, platform optimization and financial decision-making.

More from Radar