End-to-end attribution for clinics — from click to payment

What is end-to-end marketing attribution?

End-to-end attribution is a tracking setup in which every paying patient is linked to the ad source they came from. "End-to-end" means the data runs through every stage: from the ad impression to payment and repeat visits.

Standard web analytics stops at the lead: Google Analytics sees that someone submitted a form. The ad account sees that there was a click. But neither knows whether the patient showed up for the appointment or how much they paid. That data lives in the CRM or at the till.

End-to-end attribution joins these pieces on a shared key: the source tag captured on the first website visit, and the patient's phone number.

Why does a clinic need end-to-end attribution?

Without end-to-end attribution, ad decisions rely on intermediate metrics — clicks and leads — that are only loosely tied to money.

The main value is being able to cut a channel that looks good but doesn't make money, and scale a channel that looks expensive but pays off. How to calculate these numbers is covered in the guide to patient acquisition cost.

What end-to-end attribution is made of: the data chain

End-to-end attribution is a chain of six links. If one breaks, the numbers at the end are incomplete.

1. Ads. Ad links carry UTM tags; Google Ads and Meta add their own click IDs (gclid, fbclid).
2. Website. A script on the site captures the tags and landing page on the first visit and keeps them while the person thinks it over.
3. Lead. When the person submits a form or quiz, the tags go to the CRM along with their phone number and name.
4. Call. The call reaches the CRM through the telephony integration and is matched to the record by phone number.
5. Visit and payment. The receptionist books the patient, marks the visit, the payment is recorded — all on the same record.
6. Report. The CRM rolls up revenue by source; spend comes from the ad accounts.

The most common breaks happen at steps 2 and 4. Tags get lost if the person leaves the site and comes back directly a week later. Calls lose their source if the telephony isn't connected to the CRM.

Attribution models: which channel gets credit for the patient?

An attribution model is the rule for assigning a patient's revenue to a source when there were several touchpoints.

Example: someone clicked a Meta ad, left, found the clinic in Google search five days later and called.

For most clinics, first touch is a sensible starting point: it answers "which ad brought in a new patient." If the decision takes a long time (rehab, care homes, expensive treatment courses), it also helps to ask patients "how did you hear about us?" and compare their answers with the system's data.

End-to-end attribution metrics

End-to-end attribution exists to produce a handful of key metrics for each channel.

A hypothetical quarterly example (amounts in Ukrainian hryvnia, UAH):

The average ticket in this example is UAH 7,000. Meta quizzes and lead forms lead on lead volume, but Google search wins on revenue and ROMI. Lead forms have the weakest return, and if you calculate ROMI on profit rather than revenue, they may not pay off at all: optimize or cut them. Without linking leads to payments, you couldn't reach this conclusion.

How to set up end-to-end attribution at a clinic: 7 steps

1. Tag your ads. Every link gets UTM tags following one convention: source, medium, campaign. Turn on auto-tagging in Google Ads (gclid).
2. Install a tracking script on the website. It should keep the first touch for several weeks, not just the current visit.
3. Connect every form. Website forms, quizzes, and Google Ads and Meta lead forms should all send leads to the CRM with their tags.
4. Connect telephony to the CRM. Every call goes to a record by phone number. Without this, a clinic where most bookings come by phone sees only half the picture.
5. Keep bookings and payments in the CRM. The payment must land on the same record that holds the source.
6. Agree on source names. "fb," "facebook" and "FB_ads" should all become one "Facebook · ads."
7. Reconcile weekly. Spend from the ad accounts, leads, patients and revenue — in one table or one CRM report.

In 95sales CRM, source attribution works like this: the 95sales tracking code is added to the site with one line or through Google Tag Manager, remembers which ad brought the visitor (UTM tags, a Google Ads or Meta click) even if they submit the form later, and captures submissions from Contact Form 7, Elementor, WPForms and other forms. Leads from Tilda, WordPress, adsquiz and Marquiz quizzes, Google Ads lead forms and Meta (via Zapier or Make) arrive by webhook, and duplicates are detected by phone number. The Sources section turns raw tags into readable channel names.

Calls in end-to-end attribution

For clinics, calls are the weakest link in end-to-end attribution because a phone call has no UTM tag. There are three ways to tie a call to a source.

1. Matching by phone number. If the person first submitted a lead on the website (with tags) and then called, the call is matched to the same record by phone number — and the source is known.
2. Separate numbers per channel. Different phone numbers for different ad channels: a call to number X means it came from channel X.
3. Asking at the front desk. "How did you hear about us?" — with a pick list on the record. Imprecise, but better than nothing.

In 95sales CRM, calls from Binotel and Ringostat (Ukrainian telephony providers; other systems connect through a generic webhook) are matched to leads by phone number — so a call from a patient who earlier submitted a lead inherits that lead's source. More in the guide on missed calls and on the telephony integration page.

One more practical detail is repeat inquiries. If a patient from your old, archived database calls again, you want to avoid creating a duplicate with an empty source and instead bring back their existing record. In 95sales CRM, a call or lead from a number in the archive automatically returns that record to the active pipeline.

Limitations of end-to-end attribution

End-to-end attribution is never 100% accurate, and that's fine. What matters is knowing where the gaps are.

• Multiple devices. Someone sees an ad on their phone but submits the lead from a laptop — the tags may not match.
• Ad blockers and cookie consent. Some visits aren't tracked.
• Word of mouth. Recommendations from friends carry no tags; you only see them through the patient's answer.
• Offline channels. Outdoor ads and referrers are tracked through separate numbers, promo codes or the front-desk question.

What to do with end-to-end attribution data

End-to-end attribution is only useful if you make decisions with it. Weekly or monthly, look at a few questions.

1. Which channels bring patients above your acceptable CAC — cut or rework them.
2. Which channels pay off with room to spare — add budget there.
3. At which stage of the sales pipeline a given channel's leads drop off: couldn't reach them, didn't book, didn't show up.
4. Which managers handle leads from each channel best.

In the 95sales CRM analytics section, conversion and revenue by source, average ticket and LTV can be filtered by brand, website and manager — so you can answer these questions without exporting to Excel. Clinic groups and clinics running several websites should look at channels separately for each brand — see clinic chains.