What is Ads Data Hub in simple terms?

A privacy-safe query environment where you run SQL against event-level Google advertising data that Google won't let you export, and get back only aggregated results that clear a minimum-user threshold. You never see individual rows.

Is Ads Data Hub the same as BigQuery?

No. ADH runs on Google's infrastructure with strict privacy rules and returns only aggregated output; BigQuery is your own warehouse where you hold and freely query row-level data. They solve different problems.

Do I need Ads Data Hub to calculate true ROAS?

Almost certainly not. Joining ad cost with your own revenue data is a BigQuery job — you keep row-level access and none of ADH's aggregation constraints apply.

Why do my Ads Data Hub queries return no rows?

Usually because the slice is too narrow to clear the aggregation threshold. Results covering too few distinct users are filtered out by design rather than returning an error, so over-segmented queries come back empty.

Who should actually use Ads Data Hub?

Advertisers at meaningful scale who need deduplicated cross-device reach and frequency across Google properties, or custom attribution over event-level ad exposure. Below that scale, the privacy thresholds filter out most useful output.

Concepts

What is Ads Data Hub? A plain explanation of who actually needs it

TL;DR

Ads Data Hub (ADH) is often described as "BigQuery for Google's ad data", which is misleading enough to waste months. It's a privacy-safe query environment: your Google campaign data sits in Google's cloud, you send SQL to it, and you get back only aggregated results that clear a minimum-user threshold. You never see a row. That constraint is the entire product — and it means most teams asking about ADH actually want BigQuery.

The one-sentence version

Ads Data Hub lets you run SQL against event-level Google advertising data that Google will never let you export, and returns only aggregated output that passes privacy checks. You get analytical access without ever holding the underlying rows.

What the privacy model actually forces

No row-level output. Every result is aggregated. You cannot export user-level data, and you cannot look at a single user's path.
Aggregation thresholds. Results must cover enough distinct users to clear a minimum, or they're filtered out. Slice too finely and rows silently disappear rather than erroring.
Joins are constrained. Bringing your own data in is possible but governed — you can't freely join your CRM to ad logs and pull the result out.
Queries are reviewed against the rules, not just executed. A query that would leak individual users doesn't run.

This isn't friction to work around — it's the reason ADH exists. Treating the thresholds as a bug is the most common way teams end up frustrated with it.

Who genuinely needs it

Situation	ADH or not?
Cross-device reach and frequency across YouTube and Display at real scale	ADH — this is its strongest case
Deduplicated reach across Google properties	ADH
Custom attribution over event-level Google ad exposure	ADH
Joining Google Ads cost with your own revenue for true ROAS	BigQuery — far simpler, and you keep the rows
Building dashboards and recurring reporting	BigQuery
Analysing your own GA4 data in depth	BigQuery via the GA4 export
You're not spending at a scale where thresholds pass reliably	Neither — the results will be filtered out

Why most teams asking about ADH want BigQuery

The question underneath "should we use Ads Data Hub" is usually "how do we combine our ad spend with our real revenue?" That doesn't need ADH at all — it needs your ad cost and your own order or CRM data in the same warehouse, which is a BigQuery job with none of ADH's constraints. You keep row-level access, you can join whatever you like, and you can build reporting on top of it.

The distinction matters because ADH is a meaningful setup commitment. Choosing it for a job BigQuery does better costs months. We wrote a dedicated side-by-side at Ads Data Hub vs BigQuery if you're actively choosing between them.

Where an AI agent fits

To be direct about scope: Agent Planners connects BigQuery, not Ads Data Hub. What that covers is the common case — the agent writes its own SQL against your warehouse to join Google Ads, Meta and TikTok cost with your actual revenue, and turns the result into a report or a self-refreshing dashboard. If your requirement is genuinely event-level cross-device reach inside Google's privacy sandbox, that's ADH's job and you'd run it alongside, not instead.

Frequently asked questions

What is Ads Data Hub in simple terms?: A privacy-safe query environment where you run SQL against event-level Google advertising data that Google won't let you export, and get back only aggregated results that clear a minimum-user threshold. You never see individual rows.
Is Ads Data Hub the same as BigQuery?: No. ADH runs on Google's infrastructure with strict privacy rules and returns only aggregated output; BigQuery is your own warehouse where you hold and freely query row-level data. They solve different problems.
Do I need Ads Data Hub to calculate true ROAS?: Almost certainly not. Joining ad cost with your own revenue data is a BigQuery job — you keep row-level access and none of ADH's aggregation constraints apply.
Why do my Ads Data Hub queries return no rows?: Usually because the slice is too narrow to clear the aggregation threshold. Results covering too few distinct users are filtered out by design rather than returning an error, so over-segmented queries come back empty.
Who should actually use Ads Data Hub?: Advertisers at meaningful scale who need deduplicated cross-device reach and frequency across Google properties, or custom attribution over event-level ad exposure. Below that scale, the privacy thresholds filter out most useful output.