Google's Local Customer Optimization: Leave It Off For Now
Google announced a new checkbox for Performance Max on September 9 and the trade coverage treated it like a free Q4 upgrade for anyone with a physical location. Turn it on, get more nearby buyers, done. My position is the opposite for most local businesses reading this: if you cannot prove that a click turned into someone walking through your door or booking an appointment you can trace, the toggle should stay off until that measurement exists. Not because the feature is bad. Because it moves your budget toward an outcome you have no way to confirm, and you will not find out for a quarter.
The launch coverage stops at the mechanics. Search Engine Journal explains the toggle, Search Engine Land relays the announcement, Google's own documentation walks you through enabling it, and all three assume the advertiser on the other end can already see in-store results. Most local owners cannot. I have spent 10 years running Google Ads for local service businesses across the US, UK and Australia, including a London ADHD clinic I booked solid for three straight months until demand forced them to hire more specialists and outsource the overflow, and the offline measurement gap is the single most common thing I find open in those accounts. This article is about which side of that gap you are standing on, and what changes depending on the answer.
The Setting Optimizes Toward A Sale Your Account Cannot See
Brooke Osmundson reported the details for Search Engine Journal on September 9. Local Customer Optimization is a campaign-level toggle that lives only inside Performance Max campaigns using store goals. It prioritizes in-market consumers near your location using signals pulled from Maps, Waze and local Search, it was announced by Google Ads Liaison Ginny Marvin, and Google says it is "rolling out globally soon." It does not work on online-goals campaigns. It does not work on Performance Max campaigns linked to a Merchant Center feed, which means a retailer running a feed campaign has to spin up a separate store-goals campaign to use it at all.
Read that list again as a business owner rather than as a marketer. A separate campaign means a separate budget, pulled out of something that is currently working. The signals it optimizes on are location signals, which means it will find people who are physically close and physically moving. Whether those people bought anything is a question the campaign cannot answer on its own. It reports the conversion you told it to count. If the conversion you told it to count is a form fill, a phone call that nobody logs, or a "get directions" click, then the algorithm is being graded on a proxy and it will get very good at producing more of that proxy.
| Campaign type | Can you switch it on? | What it steers toward, and what you must have |
|---|---|---|
| PMax with store goals | Yes, campaign-level toggle | In-market people near your location, using Maps, Waze and local Search signals. Needs an in-store conversion you can actually verify. |
| PMax with online goals | No, not compatible | Online actions you already track in the account. Nothing changes for you right now. |
| PMax linked to a Merchant Center feed | No, not on the feed campaign | You would have to run a second store-goals campaign, funded separately, competing for the same shoppers. |
Picture a three-location dental practice in Manchester spending roughly four thousand pounds a month. They flip the toggle in October because a newsletter told them to. Bookings are taken by phone at the front desk and written into practice software that has never spoken to Google Ads. In January the account shows more conversions at a lower cost per conversion and the practice manager is pleased. Nobody in that building can tell you whether a single additional patient sat in a chair. The number went up. The chair does not know that.
A Google Release Is A Product Launch, Not A Recommendation For Your Account
Every new setting arrives wrapped in the best version of itself, because the people announcing it are the people who built it. That is normal and it is not dishonest. It just means the release notes describe what the feature does under ideal conditions, and your account is not ideal conditions. Google has no way of knowing whether your front desk logs calls, whether your point-of-sale exports transactions, or whether your CRM has an email address for every walk-in. The toggle is offered to everyone in the eligible campaign type identically.
The second reason owners get this wrong is that automation now handles so much of the account that a new automated feature feels like more of something already working. Fluency's 2026 benchmarks, drawn from a survey of more than 170 advertisers, found Performance Max drives around 45% of all Google Ads conversions and that roughly 78% of Google Ads spend now runs through Smart Bidding or Performance Max. Treat that as one survey rather than gospel, but the direction is hard to argue with: for most accounts, nearly four dollars in five is already being allocated by a machine.
When that much of the budget is machine-allocated, the conversion signal stops being a reporting detail and becomes the steering wheel. Search Engine Journal made the same point in its piece on Google automating more decisions: when conversion tracking is messy, automation scales the mess faster. Bad data used to produce a bad report. Bad data now produces a bad spending pattern, at speed, across every surface Performance Max touches, which is why handing budget control to Google's automation is a decision about your data quality before it is a decision about bidding.
The Measurement Half Of The Announcement Is The Half That Matters
Google shipped something else in the same window, and it got a fraction of the attention. Store Sales in Google Ads Data Manager lets you connect a CRM or a Google Sheet and upload real transactions so the account can match ad clicks to money taken in the building. Search Engine Journal's coverage lists what it demands: SHA-256 hashed customer data, meaning your customer emails are scrambled one way before they leave your systems so Google can match them without ever seeing the raw address; daily or weekly transaction uploads; a prior 30 days of data before dynamic conversion value works; and allowlist eligibility, so not everyone who wants it gets it. It is rolling out at the end of September 2026.
Set that requirement list next to the reality of a two-van plumbing company or a single-site physiotherapy clinic. Daily or weekly uploads of hashed customer records is not a checkbox, it is a small operational habit somebody has to own every week, forever. Most small local businesses I audit do not have a system that can produce that file, and several do not capture an email address at the point of sale at all. That is not a criticism of them. It is the actual constraint the announcement coverage skips, and it is why the same toggle can be a genuine win for a 40-location retailer and a quiet budget leak for the shop down your street.
Fewer than four ticks and the toggle is not your next move. Fixing what you count is, and that starts with the boring layer underneath: the conversion tracking hygiene most accounts quietly fail is the how-to for that repair, and this article is only the decision about whether to flip a new switch on top of it. Do the repair first. The switch will still be there.
None of this means automation is the enemy. Navah Hopkins, now Ads Liaison at Microsoft Advertising, framed the shift accurately:
Easier to take advantage of is the right description. It is also the warning. The easier a system is to switch on, the less friction there is between you and a decision you did not think through, and the more the quality of your inputs decides your outcome.
Build The Measurement Layer First, Then Flip The Switch
The order below is the one I would run for a local business, whether you do it yourself or hand it to whoever manages the account. Every step is something a non-technical owner can request, verify, or refuse.
If you are handing this to an agency, one question separates the ones who will do it properly from the ones who will just tick the box: "What in-store outcome will this campaign optimize toward, and how will we verify it happened?" A good answer names a data source. A vague answer about signals and reach is a decline. The same instinct applies to Google's own in-account recommendations to spend more, which arrive with a helpful percentage attached and no view of your margins.
What To Watch For 60 Days, And The Number That Will Fool You
Optimize toward nearby in-market users with a loose conversion definition and the account gets cheaper per conversion almost immediately, because proximity signals are abundant and the action you are counting is easy to trigger. Cost per conversion will fool you in the most pleasant direction possible. WordStream's finding that the average small business wastes about 25% of its PPC budget is not a story about bad keywords. It is a story about accounts that got very efficient at buying something nobody wanted to buy. On a five thousand dollar monthly budget that is twelve hundred and fifty dollars a month funding a graph.
Watch three things instead, and watch them for 60 days rather than 6. First, booked appointments or completed transactions counted in your own system, not in the ad platform, compared to the same period before you flipped the toggle. Second, revenue per new customer from the campaign, because a cheaper customer who spends half as much is not a win. Third, whether total volume grew or simply moved, which is the question of whether the setting found you new local demand or just relabeled the demand you were already buying. That last one is the whole argument, and it is the same question behind whether your Performance Max ROAS is incremental at all.
Sixty days is not me being cautious. Store Sales wants 30 days of prior data before dynamic values work at all, I have never seen a Performance Max change worth judging inside three weeks, and a seasonal quarter will lie to you about both. Judging this in three weeks is how a decent test gets killed and a bad one gets scaled.
Frequently Asked Questions
Should I turn on Local Customer Optimization if I don't track in-store sales yet?
No. The setting steers your budget toward people near your location who look ready to buy in person, and if you cannot confirm that anyone actually bought in person, you have no way to tell whether it worked. Your reported conversions may improve while your till stays flat. Fix what you count first: track calls through to booked or not booked, capture customer details at the point of sale, and make sure your existing conversion actions count real outcomes once each. Then enable it on one campaign and compare against your own records.
Does Local Customer Optimization work with my Merchant Center product feed campaign?
Not on that campaign. Search Engine Journal reported on September 9, 2026 that the toggle is available only on Performance Max campaigns using store goals, and it is not compatible with online-goals campaigns or with Performance Max campaigns linked to a Merchant Center feed. If you want to use it, you have to run a separate store-goals campaign alongside your feed campaign. That means a separate budget and two campaigns potentially chasing the same shoppers, so treat it as a funded test with an end date rather than a switch you flip and forget.
How long should I run it before deciding whether it worked?
Sixty days, judged against your own booking or sales records rather than the Google Ads dashboard. Store Sales needs about 30 days of prior data before dynamic conversion value works, so anything shorter mostly measures the ramp-up. Compare booked appointments or completed transactions, revenue per new customer, and whether total volume grew or just shifted from another campaign. If cost per conversion improved but your revenue did not, the setting found you cheaper clicks, not more customers.
The toggle is not really the decision. Google will keep shipping settings that allocate your money more cleverly, and each one raises the value of knowing exactly what a customer is worth to you and lowers the value of guessing. An account with clean offline data gets better every time Google automates something. An account without it gets worse, faster, for the same reason. If you want a straight read on which of those two accounts you are running before you switch anything on for Q4, book a call and we can look at it together.
