Google Ads “Report on New Customers Acquired” (August 7, 2026): Why Michigan’s Long-Cycle Businesses Will Misread the New Customer Column

On August 7, 2026, Google Ads added a third option inside Customer Acquisition settings called “Report on new customers acquired,” which unlocks the New customers and New customer value columns without applying any bidding adjustment. Search Engine Land’s Anu Adegbola reported the change the same day, crediting paid search specialist Thomas Eccel with the first sighting, and it retires a workaround advertisers had leaned on for years: selecting “Bid higher for new customers” and typing a token value of 0.01 purely to make the reporting appear.

The setting takes forty seconds to turn on. Reading the number it produces takes considerably longer — and for a large share of Michigan advertisers, the first reading will be wrong in a predictable direction.

What does “Report on new customers acquired” actually do?

It is a measurement-only switch. Selecting it exposes two columns — New customers and New customer value — and, unlike the two options above it, sends no signal into Smart Bidding. Your target CPA or ROAS behaves exactly as it did the day before.

Until this month, Customer Acquisition offered only bidding-flavored choices. Advertisers who wanted the reporting but not the behavior picked “Bid higher for new customers” and set the adjustment to 0.01 — small enough that its auction effect was assumed to be negligible. Search Engine Land credits Vasant Chaudhary with popularizing that hack. It worked, and it also left thousands of accounts carrying a bidding modification nobody actually wanted.

How does Google decide who counts as a new customer?

Two ways, and the difference between them is the entire story. If you do nothing, Google auto-detects existing customers by building a list from up to 540 days of your own tracked conversions — anyone who converted inside that window is existing, everyone else is new. If you upload a Customer Match list from your own records, Google treats anyone absent from that list as new.

Five hundred and forty days is roughly seventeen and a half months. That number is the mechanism, and once you hold it in your head you can predict how the column behaves in almost any account without running the test. Classification depends on three things: how far back your conversion tracking reaches, whether your conversion action is a purchase or a lead, and how durable the identifiers behind those conversions are once people clear cookies or switch devices.

Run that through a real business. A marine dealer in Harrison Township sells a cover and a battery in May 2024. The same customer returns in May 2026 for the seasonal refresh. Twenty-four months have passed, the 540-day window closed six months earlier, and Google files that purchase under New customers. Nothing malfunctioned — the window is simply shorter than the buying cycle. The same arithmetic hits a Rochester Hills dental practice selling crowns and any Tier 2 supplier whose purchase cycle is measured in program years rather than weeks.

The lead-gen case is worse. If your only conversion action is a form fill, auto-detection is comparing form submissions to form submissions. A fifteen-year customer who has never once filled out your web form registers as new. That is not a measurement of acquisition. It is a measurement of your form.

Does a high new-customer percentage mean acquisition is healthy?

No. On auto-detection, a high percentage is at least as likely to describe your data retention as your marketing. The column tells you how many converters Google could not find in a 540-day file — which is a different question from how many people bought from you for the first time.

Google is framing this as a transparency improvement, and on the narrow point it is right. But two things deserve pushback. First, the claim that the 0.01 adjustment had “virtually no practical impact” is a reasonable assumption nobody has published a controlled test for — an inference about a black-box bidder, not a measured result. Second, the real cost of the workaround was never the 0.01 itself. It was that the campaign carried an acquisition goal in the enabled state, so every agency handoff, account audit, and automated recommendation that reads campaign settings saw a strategic intent that did not exist.

Here is what that costs in dollars. An HVAC contractor in Livonia spending $8,000 a month turns the column on, sees 72% new customers, and concludes acquisition is working. But a furnace lasts fifteen to twenty years, maintenance renewals happen over the phone, and almost nobody in that customer base converts twice inside seventeen months. The 72% was never a rate — it was the shape of the lookback window. Six months on, that contractor has funded a campaign buying back people who already bought, at prices set for strangers.

What do the common patterns in the column mean?

Read the shape before you read the value. Five patterns cover most accounts.

• Above 90% new, no customer list uploaded. Almost certainly an artifact — thin existing-customer file, young tracking, or a buying cycle that outruns 540 days.
• 40% to 60% new with a Customer Match list active. Usually the truest reading available, and the one worth building decisions on.
• New customer value far below total conversion value. Normal where first orders are small and repeats are large. It argues for value-based bidding, not for cutting the campaign.
• New-customer count drops sharply after you upload a list. Not a decline. A correction. Do not report it upward as a loss.
• Branded search showing high new-customer share. Investigate before celebrating. Existing customers who navigate by typing your name often land in that bucket.

How should a small Michigan team switch this on this week?

Eight steps a three-person marketing team can finish in an afternoon.

1. Open each conversion-tracking campaign, go to Settings, find Customer acquisition, and record the current selection. If it reads “Bid higher for new customers” at 0.01, you are running the legacy workaround.
2. Export the last 90 days of conversions and conversion value by campaign before touching anything. That is your baseline.
3. Select “Report on new customers acquired” — in one campaign first if the account is large enough that a mistake would hurt.
4. Add the New customers and New customer value columns and save the column set so it survives your next login.
5. Upload a Customer Match list from your CRM, POS, or booking system. Include hashed phone as well as email — phone coverage is usually far better for service businesses.
6. Set a monthly reminder to refresh that list. A stale list inflates new customers every month it goes untouched.
7. Let it run 30 days before comparing new-customer CPA against blended CPA. Week-one numbers on a fresh list are not stable.
8. Do not stack this on another bidding change in the same week. One variable at a time is the only way you will know what moved.

If you run multi-location campaigns across Southfield, Sterling Heights, and Grand Rapids, upload the customer list at the account level rather than per campaign. Someone who bought from the Troy location is still an existing customer when they search near Ann Arbor, and splitting the list by branch recreates the exact misclassification you are trying to fix — a structural problem our Detroit paid search team sees constantly in accounts that have run blended CPA targets for years.

What dates should be on the calendar right now?

• August 17, 2026 — Budget-limited campaigns beating their Target CPA or ROAS begin drifting toward the target instead of outperforming it. Flip on new-customer reporting the same week and you will not know which change moved your numbers.
• September 1, 2026 — Google begins upgrading eligible Search campaigns to AI Max, changing search term matching and text customization. Establish your baseline before this lands.
• Late September through October — Michigan’s seasonal handoff. Boating and landscaping demand falls off, heating and winter service climbs. Your new-customer mix will shift for reasons unrelated to your settings.
• Early November — Refresh the Customer Match list before Q4 spend ramps. A list uploaded in August is already drifting by Black Friday.

Frequently asked questions

Will switching from the 0.01 workaround to reporting-only change my performance?

It should not, because the reporting-only option applies no bid adjustment and the workaround applied a negligible one. The honest answer is that Google has not published data either way, so export your 90-day baseline before switching. If conversions move materially in the following two weeks, you have evidence worth reporting rather than a hunch.

Do I need a Customer Match list, or is auto-detection enough?

Auto-detection is enough only if your typical repeat purchase happens well inside 540 days and your conversion tracking captures purchases rather than leads. For most Michigan service businesses, contractors, and B2B suppliers, neither condition holds, which makes an uploaded list the difference between a real number and a lookback artifact.

Does this setting work in Performance Max as well as Search?

Customer Acquisition settings live in campaigns using conversion tracking with the relevant goals, and Performance Max has carried acquisition options for some time. Check each campaign individually rather than assuming the account inherited the setting — mixed configurations are common wherever more than one person has managed the account.

Should I now bid higher for new customers since I can finally measure them?

Not until you have 30 days of clean data from an uploaded list. Bidding up for new customers before you know your true acquisition rate means paying a premium on a number that may be mostly measurement error. Measure first, decide second, and keep the two settings separate — which is the whole point of what Google shipped this month.

The short version

Turn the reporting on this week, upload a real customer list before you trust the output, and resist reallocating budget until you have a month of stable data. The businesses most likely to misread this column are exactly the ones Michigan is full of: long replacement cycles, phone-heavy sales, multiple locations, and repeat customers who look like strangers to a seventeen-month window. If you want a second set of eyes on what your column is telling you, get in touch before the September AI Max upgrades scramble the baseline.