Google Ads no longer checks Google Analytics before it uses your data
As of June 15, 2026, Google Ads stopped treating the Google Signals setting inside Google Analytics as a second gate on advertising data. Google laid out the change in a Google Analytics Help Center article, "Updates to Google Analytics Data Controls" (Google Analytics Help, 2026). Before that date, two settings had to both permit tracking before advertising cookies and user IDs reached a linked Google Ads account: Google Signals in GA4, and the ad_storage parameter in Consent Mode. After it, ad_storage is the only control that matters for Ads data. Google Signals is narrowed to Analytics-only work: linking sessions to signed-in users for behavioral reporting inside GA4.
Analytics agencies tracking the rollout, including Merkle and digital compliance vendors OneTrust and Didomi, describe the same mechanical shift: a business whose consent management platform defaults ad_storage to "granted," or does not fire the signal correctly, now has no Google Signals backstop preventing that data from reaching Ads. Before June 15, a restrictive Google Signals setting could still block advertising data even with a loose CMP configuration. That safety net is gone.
The practical change for a marketing team is narrow but not small: audit the consent management platform's ad_storage and ad_user_data signals directly, rather than trusting Google Signals settings that no longer do anything for Ads. A CMP that has been quietly misconfigured for two years will now show up as a compliance gap rather than a reporting quirk, because it is now the only thing standing between a visitor's consent choice and Google Ads' access to their data.
Databricks enters the CDP market with an agent, not a database
One day later, at its Data + AI Summit on June 16, 2026, Databricks announced CustomerLake, its first customer data platform. The product page and newsroom release describe it as an "agentic CDP" that runs natively inside the Databricks lakehouse and is governed by the company's existing Unity Catalog permissions layer, rather than as a bolt-on tool that exports data out to a separate CDP vendor. Databricks named HP, Circle K, AB InBev and Getnet by Santander as customers using the product, which is currently in private preview.
Integral Ad Science is a launch partner. According to IAS's own announcement and reporting on the launch, IAS connects a brand's first-party data to its contextual and attention measurement signals through clean-room connections inside CustomerLake, and lets brands build lookalike audiences for activation through demand-side platforms without third-party cookies. That is a direct answer to where CDPs have struggled: turning first-party data into targetable audiences once cookie-based matching degrades, without shipping raw customer records to an ad platform to do it.
The move matters less because Databricks is new to marketing software and more because of where it sits. Salesforce, Adobe and Twilio's Segment have spent the past two years pulling customer data into or alongside a data warehouse; Databricks is now building CDP functions directly on top of the warehouse itself. For a marketing team already storing customer data in Databricks for other purposes, this removes a sync step and a second permissions model. For a team on a standalone CDP, it is one more sign that "where does our customer data actually live" is becoming the more load-bearing question than "which CDP dashboard do we use."
What to do differently this quarter
These two developments land on consecutive days but share one implication: control over targeting and measurement data is consolidating into fewer, more specific checkpoints, and each checkpoint now carries more weight than before.
For the Google change, the action is immediate and mechanical: confirm the CMP fires ad_storage and ad_user_data correctly for every region with a distinct consent requirement, and stop relying on a Google Signals setting that no longer restrains Ads data collection. Teams with EU or UK traffic should treat this as a compliance item, not just a tracking one, given the removed backstop.
For the Databricks entry, there is no urgent action, but it is a reason to ask where identity resolution and clean-room matching will happen a year from now: inside a dedicated CDP, or inside whatever warehouse already holds the data. Teams evaluating a CDP purchase or renewal this year should ask any vendor, including the incumbent ones, how their roadmap responds to a warehouse-native competitor with direct clean-room partnerships already in place.
What to watch next
Databricks has not published general availability pricing or a public preview date for CustomerLake beyond its named private-preview customers; that announcement, expected once the preview period closes, will show whether this is priced as a warehouse add-on or a standalone CDP line item. On the Google side, the next signal to watch is whether Google publishes aggregate data on how much Ads traffic saw a coverage change after June 15, which would confirm how many advertisers were actually relying on the removed Google Signals backstop.
Sources: Google Analytics Help · Merkle · OneTrust · Didomi · Databricks Newsroom · Databricks Blog · Integral Ad Science · HPCwire