MarketingOS founding seats are free. Pull up a seat on Skool
Multi-location groups

You know which clinic got the booking. Do you know which ad?

Most multi-location groups can see which clinic a new patient booked at but not which campaign sent them. In a small study of 8 booking handoffs at 7 multi-location dental groups, campaign identity was lost at the booking step in 6 of the 8, and none of the 6 booking forms checked carried a campaign or click ID field. The fix is not a new platform. It is carrying the campaign through five handoffs: the ad, the location page, the scheduler, the phone, and the booked visit.

Where does the campaign get lost?

Improvado inspected booking handoffs at multi-location dental groups from public pages in mid-2026. The author calls it a small mechanism study, not an industry statistic, and the sample is dental, not therapy. The handoffs it describes are the same ones a physical therapy or chiropractic group uses, so the checklist transfers even if the counts do not.

HandoffWhat should carry throughHow it breaksHow to check it
Ad click to clinic pageCampaign tags and click ID in the URLRedirects drop the query string; one location page in the study returned an error when any query string was addedClick your own ad with a test tag and look at the address bar
Clinic page to schedulerThe same visitor ID across domainsA scheduler on another domain counts the visitor as new unless cross-domain measurement is setLook for the _gl parameter in the scheduler's URL
Scheduler or form submitCampaign and click ID saved with the bookingNone of the 6 forms in the study had hidden campaign or click ID fieldsSubmit a test booking and read what the record stores
Phone callCampaign source plus clinicOne number per clinic shows location but not campaignCall from a tracked number and see what your report shows
Booked visitThe booking matched back to the ad platformBookings never sent back, so platforms optimize to form fillsCompare platform conversions to new patients on the schedule

Breaks and counts from Improvado (Aug 21, 2026). Cross-domain behavior from Google Analytics Help. Call tracking point from Liine.

Why does the scheduler break attribution?

Google's documentation explains the mechanism. Analytics sets new cookies and IDs on each root domain, so one person moving from your site to a scheduler on another domain is counted as two users and two sessions. With cross-domain measurement configured, Analytics adds a _gl parameter to the link and the destination reuses the same IDs. Redirects that drop _gl, or navigation fired by script, break it again (Google Analytics Help).

Improvado's finding was that the breaks came from configuration, not from what the scheduling software could do: one group passed a facility code and source key into its booking app on every link.

What does that cost a six-clinic group?

Illustration: a fictional six-clinic group. Six physical therapy clinics across two metro areas, two full, one opened this year. The group spends $18,000 a month on ads, and the schedule shows 150 new patient visits booked from ads.

The ad platforms can only see the bookings that kept their campaign tag. Say 60 of the 150 did. On the platforms, cost per booked visit looks like $18,000 ÷ 60 = $300. The real figure is $18,000 ÷ 150 = $120.

The damage is uneven. The new clinic books mostly by phone because it has no reviews yet. Its schedule shows 30 new patients from ads, but only 6 carry a campaign, so its ads look like they cost $3,000 ÷ 6 = $500 per visit. A marketing head looking at that report cuts the new clinic's budget, which is the one location that needed it.

Where do groups get this wrong?

Trusting the platform's conversion count. Platforms count what they can see, usually form starts or button clicks. When the count and the schedule disagree, budgets follow the wrong one. The cost is budget moved away from clinics that are working.

One phone number per clinic and nothing else. Liine's guide says each call should carry both a campaign source and a location, and recommends a number or number pool per location. A single number per clinic answers where, not why. The cost is every phone booking landing in an unknown bucket, which hits newer clinics hardest.

Fixing tracking with condition data. Some fixes pass more detail than they should. Keep condition names and treatment details out of URLs, event names and anything sent to ad platforms; name conversions by location instead, as Catalyze Growth also recommends. Health privacy rules apply here. This is not legal advice; have your compliance lead or counsel review any change to what your site sends to ad platforms.

Which fix comes first?

Fix the handoffs in the order a patient moves through them, because each one depends on the one before.

  1. Landing pages that keep the URL. If the campaign tag does not survive the first page, nothing downstream can save it. Test every clinic page with a query string attached; the study found one that broke.
  2. Cross-domain measurement to the scheduler. Same Google tag ID on both domains, both domains listed in the tag settings, and a test click showing _gl on arrival.
  3. Hidden fields on the booking form. Save the campaign, the click ID and the clinic with each booking. This is usually a configuration change in the scheduler, not a rebuild.
  4. Calls. A tracking number per clinic at minimum, and a way to tie calls back to campaigns, so phone bookings stop landing in "unknown."
  5. Send bookings back. Only after the first four work does it make sense to send booked visits back to Google and Meta, so the platforms learn from visits rather than form starts.

What should the report look like once it works?

One row per clinic. Columns for spend, booked new patient visits from ads (from the schedule, not the platform), cost per booked visit, and the share of booked visits that carry a known campaign. That last column is the health check for the tracking itself. When it is low at one clinic, fix the handoff before you judge that clinic's ads.

The finance lead reads the cost per booked visit. The marketing head reads the known-campaign share. The operations lead reads booked visits against open slots. Same table, three readers, and nobody has to trust a platform's count to use it.

What can you check this week?

  1. Click one of your own ads per clinic with a test tag and follow it to the booking confirmation. Note where the tag disappears.
  2. Check whether your scheduler sits on another domain and whether _gl appears on the link to it.
  3. Submit one test booking and ask your scheduling vendor which fields the record stores.
  4. List the phone numbers on each clinic's ads, site and Google listing, and who owns each one.
  5. For last month, put platform-reported bookings next to new patients on the schedule, clinic by clinic.

If you want this traced for every clinic and fixed end to end, the tracking build starts with a one-week audit.

Sources

Questions groups ask

Can our scheduling software pass campaign data?

In Improvado's study, the breaks came from how booking links were set up, not from platform limits. Ask your vendor whether booking links accept and store source fields, then test one.

Do we need call tracking at every clinic?

If phone bookings are a large share of new patients, yes. A number per clinic shows location; a pool per clinic or per campaign shows which ad drove the call.

What is Google's enhanced conversions feature?

It is a Google Ads feature that sends hashed first-party data, such as an email address, to match conversions that happen after the click. Whether it fits a clinic depends on what data you are allowed to send, so have compliance review it first.

Is this a HIPAA question?

Parts of it can be, especially anything sent from pages where patients book. This article is not legal advice; involve your compliance lead before changing tracking.

Bring your numbers. We'll bring the matches.

Book a call