Customer Match in Google Ads: use the data you already hold

7 min read

Customer Match in Google Ads: use the data you already hold

7 min read

Most brands spend part of their acquisition budget on people who are already customers. They search for the brand to log in, reorder or check an order, click an ad, and the sale is counted at a new-customer cost. Customer Match in Google Ads is the feature that fixes this, using the customer lists you already hold.

The same lists do a second job. Customers who have lapsed already know the brand, and a campaign with a message written for them is a different proposition from cold acquisition, with a different target.

This article covers how we split customer data by status, what each segment gets, how we keep existing customers out of acquisition campaigns and how we measure the result on blended cost per acquisition. It should help you decide whether your CRM data is ready to use, and what to check before it goes anywhere near Google.

Why it matters to the business

When acquisition campaigns serve existing customers, two numbers go wrong at once. The budget buys sales that would probably have happened anyway, and the reported cost per new customer comes out lower than the true figure. That makes the channel look healthier than it is when leadership sets next year's budget.

Getting it right means acquisition spend goes on people who are new, reactivation runs as its own line with its own target, and the board sees one honest number: what it cost to win each new or returning customer.

It also changes what the CRM is for. Lists that sit in a database become a working input to bidding, exclusions and messaging, provided consent and data handling are in order first. The data already exists, so the work is in organising it and keeping it current, not in buying it.

How we run it

These steps run in order, because eligibility and consent decide what the later steps can do.

Check eligibility before planning. Google requires a good history of policy compliance and payment. Under the Customer Match policy, any compliant account can use customer lists for observation and exclusions, but targeting and manual bid adjustments need 90 days of account history and more than USD 50,000 of lifetime spend. That threshold decides whether a reactivation campaign can target lapsed customers directly, or only observe and exclude them.
Confirm consent and the privacy policy. Lists must be first-party data, your privacy policy must say that you share customer data with service providers, and consent is needed where the law or Google's policies require it. Google's EU user consent policy covers users in the EEA, the UK and Switzerland, so UK brands need their consent records in order before the first upload.
Split the CRM by status, using your own purchase cycle. We build at least three lists: active customers (bought within your normal repurchase interval), lapsed customers (past that interval with no order) and prospects who enquired but never bought. A brand whose customers reorder every few weeks and one whose customers buy every few years need very different cut-offs, so we set them from order data rather than round numbers. High-value customers often get a list of their own.
Sync the lists rather than uploading them once. Google now recommends Data Manager for new Customer Match connections, with connectors for sources such as BigQuery, Salesforce, Snowflake and SFTP. Membership lasts up to 540 days, a list stays eligible only while at least 100 members have been added or updated in that window, and processing can take up to 48 hours. A one-off upload leaves last year's customers excluded as if they were current, and this year's not excluded at all.
Exclude active customers from acquisition campaigns. We add the active list as an audience exclusion on non-brand Search, Shopping and Performance Max campaigns. Google announced customer list exclusions for Performance Max in January 2026. Exclusions may not apply to people who have opted out of ads personalisation, so treat this as a reduction, not a wall. We usually leave the brand campaign open, because customers searching your name to log in or reorder should still find you.
Tell the bidding who is new. The new customer acquisition goal has a mode that bids higher for new customers and one that bids only for them, which Google suggests pairing with a separate campaign for existing customers. Existing customers are defined from your lists and tag, and Google warns that a list with fewer than 1,000 members may hurt the campaign's ability to acquire new customers.
Write a separate message for lapsed customers. Lapsed customers get their own campaign and copy: what has changed since they last bought, the new range, a reason to come back. Performance Max has a re-engagement goal for lapsed customers you define and upload through Customer Match. In Search, an account that meets the targeting threshold can run a lapsed-only campaign, and one that does not can still add the list in observation to see how lapsed customers behave. This is closer to remarketing than acquisition, but built on purchase history rather than site visits.
Close the loop with enhanced conversions. Enhanced conversions sends hashed customer data with each conversion so Google can match it to signed-in accounts, which improves measurement accuracy and gives the new customer goal a better read on who actually converted. For lead generation, enhanced conversions for leads ties CRM outcomes back to the campaign that produced them.

How to measure it

We measure on blended CPA: total paid search spend divided by the total number of new and reactivated customers it produced, with each group also shown separately. It is the figure that stops a cheap acquisition CPA being flattered by sales to people you already had.

New customer CPA from acquisition campaigns, using the new and existing customer split in the customer acquisition reporting rather than total conversions.
Share of acquisition conversions that come from existing customers. Once exclusions are live this should fall. If it does not, check list freshness and match rate.
Reactivation rate and revenue from the lapsed segment, against the cost of the reactivation campaign.
Match rate on each upload, shown once processing finishes. A sudden drop usually points to a change in the format of the CRM export.

Testing and reporting upwards

To know whether reactivation adds sales rather than catching customers who would have come back anyway, hold back a random part of the lapsed list from the campaign and compare return rates between the two groups. Our note on incrementality testing explains the method.

Report blended CPA, new customers won, lapsed customers reactivated and the revenue from each. That gives leadership one view of customer cost across acquisition and retention, instead of two campaigns claiming the same sale.

When it does not fit, and the mistakes we see most

Small or stale lists. A list that drops below 100 refreshed members stops being eligible, and lists under 1,000 limit the new customer goal. Brands with few customers get more from site-based audiences.
Accounts under the spend threshold. Without targeting access, lists can only observe and exclude. That is still worth doing, but a lapsed-only Search campaign has to wait.
Consent gaps. If you cannot show how the data was collected and that sharing it was disclosed, it should not be uploaded.
Work email lists. In our B2B audits, lists made up mostly of work addresses usually match a smaller share of people, so we check the match rate before building a plan on them.
Excluding customers from brand search. Blocking your own customers from the brand campaign sends them to the organic listing at best, and to a competitor's ad on your name at worst.
Discounting by default. Lapsed customers do not always need an offer. Test a message about what has changed before giving margin away.

Questions to ask your team or agency

Our paid search team sets this up as part of account management, from list design and syncing through to blended CPA reporting. Whoever runs yours, these questions will show where it stands.

Is our account eligible for Customer Match targeting, or only for observation and exclusions?
How are customer lists synced, and when was each one last refreshed?
What share of conversions in our acquisition campaigns comes from existing customers?
How do we define a lapsed customer, and is that based on our own repurchase data?
Who has checked that our consent records and privacy policy cover this use?

We use Customer Match in both e-commerce PPC and B2B PPC accounts, with the lists split by what each customer group is worth.

Frequently asked questions

What is Customer Match in Google Ads?

Customer Match lets advertisers upload first-party customer data, such as email addresses and phone numbers collected with consent, so Google can match it to signed-in users. The lists can be used to exclude existing customers from acquisition campaigns, observe how customers behave, target them directly in eligible accounts, and tell Smart Bidding which conversions come from new rather than returning customers.

Who is eligible for Customer Match in Google Ads?

Under Google's Customer Match policy, an account needs a good history of policy compliance and payment. Every compliant account can use customer lists for observation and exclusions. Using lists for targeting and manual bid adjustments also needs at least 90 days of Google Ads history and more than USD 50,000 of lifetime spend, with other currencies converted at the average monthly rate.

How often should Customer Match lists be updated?

As often as your customer data changes, ideally through an automatic sync rather than manual uploads. Membership lasts up to 540 days, and a list stays eligible only while at least 100 members have been added or updated in that window. A stale list excludes people who are no longer customers and misses new ones, which weakens both exclusions and new customer bidding.