Google Ads Is Auto-Labeling Your Customer Lists Now: What Changed August 18 and What to Fix
Written by
Aerin Kim
Since August 18, 2026, Google Ads auto-labels every unclassified conversion-based customer list. Here is what changed, why it affects Smart Bidding, and how to audit your lists now.
If you have not opened Audience Manager in a while, some of your customer lists may already carry a label you never assigned. Starting August 18, 2026, Google Ads began automatically classifying every eligible conversion-based customer list that an advertiser has not manually labeled, sorting each one into one of three lifecycle buckets: existing customers, new customers, or other customer segments. Google started activating the feature for eligible accounts back on June 17, 2026, but the actual data processing, the point where labels get assigned to real lists, only began on August 18.
This is a genuinely quiet rollout. There was no major Google Ads blog post or press announcement driving the news the way a Performance Max update usually gets one. Strategists first spotted it through in-product notifications inside their own accounts, and coverage spread from there. That matters for how you should treat this post: it is not optional context you can skim past, it is a change that may have already reclassified real lists in your account without you clicking anything.
This post covers exactly what changed, why it feeds directly into Smart Bidding and Performance Max, and the specific steps to audit your account this week. If you are running acquisition campaigns and building out creative for them at the same time, the AI Image Generator in Miraflow AI is useful for producing the product and lifestyle visuals a customer acquisition campaign needs without a full photo shoot, and the YouTube Thumbnail Maker covers the same need for any remarketing creative that runs alongside your video campaigns.
What Actually Changed on August 18
Google now automatically assigns a customer type to any eligible conversion-based customer list that you have not classified yourself. There are three broad buckets:
- Existing customers
- New customers
- Other customer segments
Underneath those three broad buckets sits a more granular, nine-label taxonomy Google uses to actually sort lists: All customers, Cart abandoners, Converted leads, Disengaged customers, High value customers, Loyalty program members, Paid subscribers, Purchasers, and Qualified leads. A list you built months ago for a completely different purpose, say, everyone who added an item to cart in the last 90 days, may now be sitting under a label like Cart abandoners or Purchasers that Google inferred from your conversion data, not one you chose.
| Broad Customer Type | Granular Labels Underneath It |
|---|---|
| Existing customers | All customers, Purchasers, High value customers, Loyalty program members, Paid subscribers |
| New customers | Converted leads, Qualified leads |
| Other customer segments | Cart abandoners, Disengaged customers |
If you would rather keep making that call yourself, Google gave advertisers a way out, but the window mattered: disabling conversion-based customer list classification in Account Settings before August 18 prevented the auto-labeling from touching your lists at all. If you are reading this after that date and never touched the setting, assume auto-labeling has already run across your eligible lists and start with the audit steps below rather than the opt-out.
Why This Is Not Just a Labeling Cosmetic Change
The reason this deserves real attention instead of a shrug is that customer type labels are not cosmetic metadata sitting off to the side. They feed directly into two of Google Ads' most consequential systems: Smart Bidding and Performance Max targeting.
Google's own customer lifecycle goals, specifically New Customer Acquisition and Re-engagement bidding, rely on accurate customer type labels to function correctly. When a list mixes new and returning customers together under one label, or gets auto-classified into a bucket that does not match how you actually use it, you lose the ability to bid differently for acquisition versus retention and you lose the ability to exclude the right audience from the right campaign. Performance Max makes this risk larger, not smaller, since Performance Max campaigns already drive a significant share of total Google Ads conversions for most advertisers, and Performance Max's own audience and data exclusion settings, covered in more depth in our breakdown of Performance Max's 2026 optimization controls, depend on exactly this kind of clean customer list segmentation to work as intended.
Put plainly: a mislabeled list does not just look wrong in a report. It can quietly cause Smart Bidding to treat an acquisition campaign as if it already has retention signal mixed in, or cause a re-engagement campaign to bid for people who were never actually existing customers to begin with. Neither failure throws an error. Both just show up later as underperformance that is hard to trace back to its actual cause without checking the labels directly.
How to Audit Your Account This Week
There is no bulk, cross-account tool for this, which means the audit is manual and per-account. Budget real time for it if you manage more than a handful of lists.
- Go to Tools and Settings, then Shared Library, then Audience Manager. This is where every customer list in the account lives, along with whatever customer type label is currently attached to it.
- Open each conversion-based list and check its assigned label against what the list actually contains. Pay closest attention to any list built before this feature existed, since those are the ones most likely to now carry a label nobody chose.
- Split any list that mixes new and returning customers into two separate lists. A single list serving both purposes is exactly the scenario that breaks lifecycle bidding goals, whether the label was auto-assigned or not.
- Apply the correct label deliberately to each list, rather than leaving Google's inference in place by default, even if the inferred label happens to be correct today. Deliberate labeling is more durable as your list definitions evolve.
- Cross-check any list feeding a Performance Max campaign's audience signals or data exclusions specifically, since that is where a mislabeled list has the most immediate, measurable effect on spend.
If your organization runs the same customer list logic across multiple accounts, plan for this to take meaningfully longer, since each account needs its own manual pass through Audience Manager rather than one shared fix.
What to Do If You Disabled Auto-Labeling
If you turned off conversion-based customer list classification before August 18, your lists were not touched by this rollout, but that does not mean the work is done. Unlabeled lists still cannot participate in customer lifecycle bidding goals the way labeled ones can, so opting out trades away auto-labeling risk for a different tradeoff: you now need to manually label every list you want Smart Bidding and Performance Max to treat with lifecycle awareness, on your own schedule instead of Google's. Treat the opt-out as a deferral, not a fix, and put manual labeling on your team's roadmap rather than letting it sit indefinitely.
If your acquisition and re-engagement creative is still being produced separately by hand, our guide to AI UGC ad prompts for product and lifestyle content covers a faster way to generate distinct creative sets for each customer segment once your lists are correctly split.
Common Mistakes With the New Auto-Labeling
- Assuming a label that looks correct today will stay correct as a list's membership shifts over time. Auto-labeling is not continuously re-evaluated the way a dynamic segment would be, so review labels periodically rather than treating a single check as permanent.
- Leaving mixed new-and-existing-customer lists unsplit because the auto-assigned label technically applies to most of the list. Most is not all, and the minority membership is exactly what corrupts lifecycle bidding signal.
- Treating this as a Performance Max-only concern. Standard Search and Shopping campaigns using customer match audiences are affected by the same labels.
- Forgetting to check lists that feed audience exclusions rather than audience targeting, since an incorrectly labeled exclusion list can quietly let in exactly the audience you meant to keep out.
- Waiting for a formal Google announcement before acting. As this rollout shows, Google does not always pair a meaningful account-level change with a prominent blog post, and the account itself is the more reliable source of truth.
Where This Fits Alongside Miraflow AI
Customer list hygiene determines whether your bidding signal is clean, but it does not fix a creative shortage once your acquisition and re-engagement campaigns are correctly split. The AI Image Generator in Miraflow AI supports text-to-image generation, image-to-image editing, and inpainting, useful for producing distinct creative sets for new-customer acquisition campaigns versus re-engagement campaigns without a separate photo shoot for each. The cinematic AI video generator covers the matching video formats for the same split. If you are building out full asset groups for Performance Max specifically, our guide to Performance Max's 2026 controls and our explainer on Google's target-based bid strategy change from August 17 are both directly relevant companion reads, since all three changes landed inside the same two-week window and compound on each other. Our look at Google's Ask Advisor AI agent is also worth reading if you want to see how Google is using AI to help advertisers navigate exactly this kind of control sprawl. Browse more marketing-focused guides on the Miraflow AI blog.
Frequently Asked Questions
Did this affect every customer list in my account? Only conversion-based customer lists that were eligible and that you had not already manually classified. Lists you had already labeled yourself were left alone.
Can I still opt out now that data processing has started? The documented opt-out window was before August 18, 2026. If you are past that date, treat your eligible lists as already processed and focus on the audit steps above rather than trying to retroactively disable the feature.
Does this apply to Performance Max specifically, or all campaign types? Customer type labels apply wherever customer match and lifecycle bidding goals are used, which includes Search, Shopping, and Display in addition to Performance Max. Performance Max is simply the campaign type where the downside of a mislabeled list tends to be largest, given how much conversion volume it typically drives.
What is the difference between the three broad labels and the nine granular ones? Existing customers, new customers, and other customer segments are the three high-level buckets that lifecycle bidding goals actually key off of. The nine granular labels, like Cart abandoners or Loyalty program members, are a more specific classification layer Google uses underneath those three broad buckets.
Is there a way to check every account at once if I manage several? Not currently. The audit has to be run manually inside Audience Manager for each account individually, so budget proportionally more time if you manage a multi-account setup.
Why did Google not send a prominent announcement about this? Google Ads regularly rolls out account-level and infrastructure changes without a dedicated blog post, especially changes framed as improving consistency of existing systems rather than adding a new advertiser-facing feature. This is a useful reminder to check Audience Manager and account notifications periodically rather than waiting for headline coverage.
Conclusion
This is one of those changes that is easy to miss entirely and expensive to leave unchecked. Google auto-labeling unclassified customer lists sounds like a small housekeeping update, but because those labels feed directly into Smart Bidding's lifecycle goals and Performance Max's audience exclusions, a mislabeled or auto-assigned list can quietly cost you acquisition efficiency or let excluded audiences back into a campaign without ever throwing an obvious error. The fix is not complicated, open Audience Manager, check every conversion-based list against what it actually contains, split anything mixed, and label deliberately going forward, but it is manual, per-account, and easy to keep postponing. Do not.
