Google Ads Is Deleting Your Historical Data. Here’s What You Need to Do Before It’s Gone.

The new data retention policy starting June 1 — what’s changing, what’s at risk, and the exact steps to protect your performance history.

Starting June 1, 2026, Google Ads is changing how long it retains historical performance data. The change is quiet in its communication and significant in its consequences — particularly for brands and agencies that have been treating Google Ads as the default home for their performance history.

This article explains exactly what’s changing, which data categories are affected and over what timelines, which types of advertisers face the most meaningful risk, and what to do right now to prevent years of accumulated performance data from becoming permanently inaccessible.

What’s Actually Changing in Google’s Data Retention

Google Ads is implementing a tiered retention system that applies different storage windows depending on the granularity and type of data:

Granular performance data (hourly, daily, weekly): 37 months

The data that most campaign managers use for day-to-day and month-over-month performance analysis — impression counts, click data, conversion metrics, cost data at the daily and weekly level — will only be accessible for 37 months (approximately three years and one month) from the date it was generated.

After this window, the data is deleted from the Google Ads interface and is no longer retrievable through the Google Ads API. There is no archive. There is no recovery process. Once the 37-month window passes, the data is gone.

Aggregated performance data (monthly, quarterly, annual): 11 years

Summarised data at the monthly, quarterly, and annual levels remains available for up to 11 years. This means that high-level performance trends and year-over-year summaries will still be accessible, but the granular daily and weekly data that underpins them will not.

Audience and reach metrics: 3 years

Specific audience data — unique users, impression frequency, frequency distribution — has a shorter retention window of three years. This is the data most relevant to frequency cap management, audience saturation analysis, and reach campaign performance evaluation.

Why This Matters More Than It Appears

The immediate reaction from many advertisers will be: “I can still see monthly and annual data, so I’m fine.” This reaction underestimates what the loss of granular data actually means in practice.

• Year-over-year seasonal analysis requires daily data. Comparing this Q4’s daily performance curve against Q4 from three years ago — to understand whether a specific week’s performance anomaly is seasonal or campaign-specific — requires daily granularity. Monthly averages smooth out the patterns that daily data reveals. Once the 37-month window closes on daily data, this level of seasonal analysis becomes impossible for older periods.

• Frequency cap optimisation requires granular reach data. Understanding how impression frequency affects conversion rate — the curve of diminishing returns at which additional impressions start hurting performance — requires impression frequency data at sufficient granularity to see the pattern. The 3-year window on this data means that frequency learning from campaigns run more than three years ago becomes inaccessible.

• Agency-client transitions lose historical context. When an agency takes over a Google Ads account, one of the most valuable inputs to their initial strategy is the account’s full performance history — the campaigns that worked, the ones that didn’t, the seasonal patterns, the audience performance data. With a 37-month rolling window, accounts that have been managed for many years will progressively lose the institutional memory that sits in the platform.

• Benchmarking requires multi-year baselines. CAC benchmarks, ROAS baselines, and cost-per-lead targets that are grounded in historical performance data become less reliable when the granular historical data that validated them becomes inaccessible. The benchmarks remain — in documents, in annual reports — but the ability to audit and verify them against the source data disappears.

Who Faces the Most Risk

Not all advertisers are equally affected. The advertisers with the most to lose from this policy change share specific characteristics:

• Seasonal businesses with annual peak periods. Brands whose Google Ads performance has distinct annual peaks — retail at holiday season, travel at summer, tax services in Q1 — use multi-year daily data to model expected performance and set appropriate targets. Three years of data is barely sufficient to establish reliable seasonal baselines; beyond three years, the patterns become more robust. Losing access to daily data beyond 37 months removes the deeper seasonal context that mature performance models depend on.

• Agencies managing long-standing client accounts. Agencies that have managed client accounts for four or more years have built performance context that exists in the platform’s historical data. When this data rolls out of the 37-month window, the account’s history becomes reconstructed from documents and reports rather than queryable from the source.

• Brands that have never exported or warehoused their Google Ads data. The majority of Google Ads accounts have no independent data warehouse — the performance data lives in Google Ads and is accessed through the interface or API when needed. For these accounts, the June 1 policy means that any daily or weekly data currently in the account that is more than 37 months old will be deleted, and any data that reaches the 37-month threshold after June 1 will be deleted on a rolling basis.

• Performance marketers doing attribution and incrementality work. Granular historical data is essential for building attribution models, conducting incrementality analysis, and establishing the counterfactuals that make rigorous performance measurement possible. The 37-month window limits how far back this analysis can reach in the Google Ads source data.

The Immediate Action Plan to Do Before June 1, 2026

The response to this policy change is straightforward in concept and requires genuine operational investment to execute properly. Here’s the priority sequence:

Step 1: Audit What You Currently Have

Before exporting anything, understand what historical data currently exists in the account and which portions are most valuable. For most accounts, the priority data is:

• Daily performance data at the campaign and ad group level for the past 4+ years
• Keyword-level daily performance data for core campaigns
• Audience segment performance data (particularly if running audience-targeted campaigns with frequency data)
• Ad creative performance data with daily granularity

Log into Google Ads and check the date range available for your most important reporting segments. Any data beyond 37 months from today is at risk under the new policy.

Step 2: Export Historical Data to a Permanent Storage Location

The export can take several forms depending on the organisation’s technical infrastructure and resources:

• Manual exports for smaller accounts: Google Ads allows direct CSV export of reports at configurable date ranges. For accounts with manageable data volume, a systematic manual export of key reports — downloading daily data in 12-month batches going back as far as available — creates a local archive of the historical performance record.

• Google Ads API for larger accounts: For accounts with significant data volume, the Google Ads API provides programmatic access to performance data. Setting up an API-based export that pulls and stores historical data before the deletion window, and then runs on a regular schedule to archive data before it ages past 37 months, is the robust long-term solution.

• Google Ads Data Transfer (BigQuery): Google offers a managed data transfer service that automatically syncs Google Ads data to BigQuery on a daily basis. Setting this up before June 1 means that all new data generated after setup is automatically warehoused, and historical exports can populate the same BigQuery dataset for older data.

• Third-party analytics and BI tools: Platforms like Supermetrics, Funnel.io, or Looker Studio with appropriate connectors can pull Google Ads data into third-party storage on a scheduled basis. If the organisation already uses one of these platforms, configuring a historical backfill export before June 1 is the path of least resistance.

Step 3: Establish an Ongoing Archival Process

The June 1 policy makes a one-time data export insufficient. The 37-month window is rolling — data that is currently within the window will eventually age out of it unless it is archived. The solution is an ongoing archival process that regularly exports data to permanent storage before the retention window closes.

At minimum, this means: monthly exports of the previous month’s daily performance data to a permanent location. A disciplined monthly export practice means that no data is ever lost because it was simply not exported before the rolling window closed on it.

Step 4: Document Audit Trails for Compliance-Sensitive Accounts

For brands in regulated industries where advertising performance records have compliance implications — financial services, healthcare, legal — the data retention policy change requires explicit documentation of what data is being retained, where, and for how long. Standard compliance processes that relied on Google Ads as the authoritative data source need to be updated to reflect the new retention limits and the organisation’s own archival process.

Long-Term Data Infrastructure Implication

Google’s data retention policy change is a direct consequence of a broader trend: advertising platforms are increasingly managing their data infrastructure in ways that serve their own operational needs rather than advertisers’ historical access needs. This is not a criticism — it’s an observation about where data sovereignty sits in the current ecosystem.

The brands that have built independent data infrastructure — their own BigQuery instances, their own performance databases, their own analytical layer separate from the platforms they use — are insulated from this kind of policy change. Their data lives in their environment, under their control, accessible on their timeline. Platform policy changes affect their access to new data from the platform, not their access to the historical data they’ve already archived.

The brands that treat advertising platforms as their data warehouses are perpetually at the mercy of platform data policies. Google’s retention change is a concrete example of why this dependency is a risk.

The right long-term response to this policy is not just an emergency export before June 1. It’s the beginning of a data ownership posture — building the infrastructure to control your own performance history independent of the platforms that generate it.

What to Do If You Miss the Window

If June 1 has already passed and daily performance data older than 37 months has been deleted before an export was completed, the options are limited but not non-existent:

• Reconstruct from existing reports. If the account has existing saved reports or scheduled report emails that captured historical data before deletion, these become the archival source. They may not have daily granularity, but they may have weekly or monthly data that provides sufficient context for most planning purposes.

• Use aggregated data as the baseline. Monthly and annual data retained for 11 years can serve as the baseline for trend analysis, seasonal modelling, and benchmark setting, even without daily granularity. It’s a reduced capability, but it’s not zero capability.

• Supplement with external data sources. Third-party competitive benchmarking tools (SEMrush, SpyFu, SimilarWeb) retain historical data on their own schedules and can provide category and competitor context that partially compensates for the loss of granular first-party Google Ads data.

The better path is not to be in this position. The data export process described above is not difficult or expensive — it requires operational attention, not significant investment. The cost of losing years of granular performance data far exceeds the cost of the export process that preserves it.

Conclusion

The Google Ads data retention policy change is a reminder that data stored in third-party platforms is stored at the platform’s discretion, not yours. Access that exists today may not exist tomorrow — not because of malicious intent, but because platforms manage their infrastructure for their own purposes.

The response to this is ownership: taking the data that your campaigns generated and storing it somewhere you control, on terms you define. For most brands, the June 1 deadline is the immediate urgency. The longer-term opportunity is building the data infrastructure that makes this kind of platform policy change irrelevant.

The Brisk Digital helps brands build data infrastructure for their paid media programmes — from Google Ads export pipelines to BigQuery warehousing to the analytical layer that makes historical data useful. If you want help setting up a data archival process before the June 1 cutoff, we’re here.