Customer Match: upload your whole CRM or only the recent customers?

The usual advice is to upload your full customer list, the bigger the better, but I don't think that holds up everywhere. Take a dental clinic with 12,000 past patients, most of whom haven't visited in four years. If you feed all of them in, Google builds lookalikes partly from people who already churned, and that muddies the signal.

My lean is to upload only the segment you'd actually want more of, like patients from the last 18 months or people who booked a high-value treatment. The catch is that match rates drop with a smaller list, and a very small one may be too thin to do anything useful. So where do you draw that line? I'm also not sure match rate ends up mattering more than list quality.

Depends what the upload is for, so settle that first. If it's a seed for lookalikes, Elena's right that long-churned patients dilute the signal, and a tight recent or high-value segment makes more sense. But if the same list is also used to exclude current patients from prospecting, you want the whole thing, or you end up paying to show acquisition ads to people who are already in the chair. Splitting it works well: recent segment as the seed, full list as the exclusion.

Check that the recent segment will actually match before you split the list. Google generally wants around 1,000 matched users before a Customer Match list is useful for targeting, and email alone often matches only about half the rows. So a dental clinic with 3,000 patients from the last 18 months could easily end up with just a few hundred matches. Adding phone numbers and name plus postcode as extra identifiers is usually what gets the seed over the line, so don't upload email only.

Agreed on the identifiers. The 1,000 figure is Google's threshold, though, and minimums vary by platform. LinkedIn Matched Audiences has historically needed around 300 matched contacts, so a seed that's too thin for Google could still work there. Check the current minimum in each platform's docs before you split the list.

Whether the list fits the platform matters more than hitting the threshold. LinkedIn matches on work email and professional profile data, so a dental clinic's patient list will mostly come back unmatched there, even if it clears 300 on paper. Patient records tend to be personal emails and phone numbers, and most of those people don't have a usable LinkedIn profile. The 300 figure only gets realistic when the seed is a B2B list, like past clients at a software company. Which kind of list are you working with?

Agreed on LinkedIn. For a clinic, what matters is whether the seed beats a broad audience on bookings, not how many people it matches. And that only works if bookings actually reach the platform as a conversion. Are they tracked there, or only in the practice software?

I'd guess most clinics land in the second group. Practice software knows who booked, but the ad platform only finds out if the booking gets sent back, either through a confirmation tag on a thank-you page or an offline conversion upload. Without that, the campaign optimizes for form fills or calls, which can look fine in the dashboard while actual bookings barely move. Honestly, for a clinic the creative still does most of the work. The seed just decides who sees it first. Which booking system are they on?

Before the booking system question even matters, I'd check whether a patient list can go into an ad platform at all. Being a patient reveals health information, which can fall under the sensitive-data rules in Google's and Microsoft's customer match policies, and under special-category rules in GDPR. Some clinics upload anyway and only find out later, when the list gets disapproved or someone questions the consent basis. Often the cleaner route is conversion tracking with broad targeting and letting the creative do the work. Does the clinic have marketing consent on file for those patients, or is that still an open question?