What is patient-journey attribution?
Last reviewed
Patient-journey attribution is the practice of linking the marketing touches a patient had, such as a search, an ad, a WhatsApp chat or a referral, to what actually happened next: an appointment, a visit, a treatment and revenue. Unlike lead attribution, it gives credit only when a real patient is treated, which is the only measure a hospital board trusts.
Updated 24 October 2026: merged in our guide on what attribution can and cannot tell you, and added platform bias, DPDP limits, holdout tests and a harvest, capture and create budget split.
New to the topic? Start here: Healthcare digital growth sets out the reading order.
The short definition
Patient-journey attribution answers one question: which channels and campaigns produced patients who were actually treated, and at what cost. It follows a person from first touch to billing, across channels and units, using the hospital’s own records rather than an ad platform’s estimate.
It is different from the attribution most marketing teams report. Ad platforms attribute conversions to their own clicks and count form fills or calls. A hospital needs to know which of those became patients, and that data lives in the CRM and the hospital information system, not in the ad account.
How it works
| Step | What happens | Where the data lives |
|---|---|---|
| Touch | Patient sees an ad, searches, reads a page, gets a referral | Ad platforms, analytics, referral records |
| Enquiry | Patient calls, fills a form, messages on WhatsApp or walks in | Telephony, website, WhatsApp, front desk |
| Identity | Enquiry is matched to a person, usually by mobile number | CRM |
| Appointment | A slot is booked and then honoured or missed | CRM and appointment system |
| Treatment and billing | The visit, admission or procedure is billed | Hospital information system |
| Credit | Treated patients and revenue are credited back to the source | CRM reporting |
The models, in plain language
- First touch: all credit to the channel that started the journey. Good for judging awareness spend.
- Last touch: all credit to the final channel before the enquiry. Simple, and the default in most ad platforms, but it over-credits search and brand.
- Multi-touch: credit shared across touches. Closer to reality but needs clean data across channels.
- Source of enquiry: the practical starting point for most hospitals. Record how each enquiry arrived, then follow it to treatment.
Most Indian hospital groups should start with source of enquiry, linked all the way to treatment. It is simple, defensible, and it already answers the board’s real question. Multi-touch models can come later, once the data is clean.
Why platform numbers lean one way
Platform-reported conversions are not random noise. They are biased, consistently, in a direction that flatters the platform. Last-click over-credits search, and brand search most of all, because it harvests demand that other things created. View-through and engaged-view conversions over-credit video. A model built inside one platform cannot see the others at all, so the sum of every platform’s claimed conversions routinely exceeds your actual appointment count, sometimes by a wide margin.
This matters operationally because the over-credited channels are also the easiest to scale. Naive optimisation concentrates spend into brand search and remarketing, the two places where incrementality is lowest. You will see improving cost per acquisition and flat patient volume for two quarters before anyone connects the two.
What you need to set it up
- One source taxonomy used by every channel and unit.
- A CRM that captures every enquiry, including calls and WhatsApp.
- A patient identity rule, usually the mobile number plus name, to join records.
- A link from the CRM to appointment and billing data.
- Consent recorded for how the data is used, under the DPDP Act.
The rest of this guide covers what you can and cannot know from this data. I have written about the CRM foundations in hospital CRM implementation.
Privacy and DPDP change the method, not the ambition
Health data is sensitive in every framework, and India’s regime is explicit about purpose limitation and consent. In practice three constraints bind you.
- You cannot build audiences from clinical attributes without a consent basis that specifically covers it. “They were our patient” does not cover advertising.
- Condition-level identifiers should not go into advertising platforms at all, whatever a vendor says is technically possible. The reputational exposure of an oncology audience leak is not worth any amount of efficiency.
- Consent withdrawal has to reach every outbound system, so audience lists need rebuilding on a cycle rather than accumulating.
What this removes is the richest form of targeting. What it leaves is still workable: contextual and intent signals, locality, non-clinical behavioural segments and aggregate measurement. The groups that handle this well treat the constraint as a design input from day one. The ones that handle it badly build on patient-attribute targeting and then dismantle it, usually after a legal review rather than before.
Where it breaks
- Walk-ins and phone calls that never get a source recorded.
- Family members enquiring on behalf of the patient, under a different number.
- Doctor referrals logged at the front desk but not in the CRM.
- Long gaps between enquiry and treatment, especially for elective procedures.
- Units that define sources differently, so group totals cannot be compared.
None of these are reasons not to start. Even partial attribution to treated patients beats perfect attribution to leads.
A typical journey shows why. A woman sees a hoarding near the flyover in March. In June her father is diagnosed with something that needs a surgeon. She searches on her phone, reads three pages, asks a cousin, calls a number she found on a listing site rather than yours, and walks in on Thursday with a referral letter from a GP she had already visited in May. Her registration form records the source as “walk-in”. Every attribution model will get this wrong. The useful question is how much wrongness you can tolerate and still allocate budget sensibly.
The holes that are hardest to close
- Walk-in as a source. The largest single category in most units, and it means “the front office did not ask” about as often as it means anything else.
- The contact centre gap. A campaign call that lands on a published number rather than a tracked one is untraceable the moment it is answered, and patients often call the number they already have.
- Attribution windows versus decision cycles. Cookie windows run for days and weeks. Oncology and joint replacement decisions take months, so the window expires before the decision is made.
- Shared devices. One household, one phone, three patients. Platform identity graphs were not built for this.
- Identity matching at billing. Shared numbers, the relative who registers on the patient’s behalf and the number typed in wrong all break the join from enquiry to bill.
What to report
Report a small set: treated patients and revenue by source, cost per treated patient by channel, and the share of treated patients whose source is unknown. That last number tells you how far to trust the rest.
Three numbers I trust while the join is incomplete
Until enquiries link cleanly to billing, I lean on three numbers and treat everything else as directional evidence rather than fact.
- Identified enquiries, deduplicated. A contactable person who asked about a service, with duplicates collapsed by phone number across form, call, chat and walk-in. This is almost always lower than the sum of what each channel claims, and the gap is a useful hygiene metric in itself.
- Enquiry-to-appointment and appointment-to-consult. Internal, unaffected by platform politics, and where most of the real loss lives. You do not need attribution to measure these rates, and improving them raises the yield of every channel at once.
- Total volume against total spend. At unit and service-line level, over a rolling quarter. If spend rose and new patients in that service line did not, the channel reports are telling you a story about credit, not causation.
Who is asking for the number, and what they actually want
When someone demands attribution they are rarely asking a measurement question. The answer changes with the person asking.
- The chief executive asking “is digital working?” wants reassurance that spend is not waste and a basis for defending it. A total-volume-against-total-spend trend by unit answers that better than any channel report.
- The chief financial officer wants the marginal cost of the next patient and whether it is rising. Give cost per identified enquiry and the enquiry-to-appointment rate, and be honest that the product of the two is an estimate.
- The unit head wants to know why footfall is flat while the group celebrates a campaign. The honest answer usually involves slot availability or a consultant’s calendar, and attribution is the wrong instrument.
- The agency wants credit. That is a legitimate commercial interest and a poor basis for your allocation.
I have made the mistake of answering all four with the same deck. It satisfied nobody and trained the organisation to distrust the numbers, because each audience could see the part that did not address them.
A worked example
A patient’s daughter searches for a cardiologist in the evening, clicks a search ad, reads two doctor pages and sends a WhatsApp message. The contact centre replies the next morning and books a consultation. The patient attends, has tests, and a week later returns for a procedure, paid partly by insurance.
| What each system sees | What it credits |
|---|---|
| Ad platform | A click and possibly a conversion, if the WhatsApp tap was tracked |
| Website analytics | A session from paid search with two page views |
| CRM without integration | A WhatsApp enquiry with no source, under the daughter’s number |
| CRM with a source rule and identity match | A paid search enquiry, linked to the patient, the consultation and the procedure |
| Billing | Two bills: consultation and procedure, with the insurer as payer |
Only the last two rows, joined together, tell the hospital that a paid search campaign produced a treated cardiac patient. That join is what patient-journey attribution means in practice.
A simple maturity path
- Level one: every enquiry has a source recorded, including calls and WhatsApp.
- Level two: enquiries are linked to appointments, so you can report booked and honoured appointments by source.
- Level three: appointments are linked to billing, so you can report treated patients and revenue by source.
- Level four: multiple touches are recorded, such as a first search and a later referral, and credit is shared.
Most hospitals get the biggest value from moving from level one to level three. Level four is useful mainly for large marketing budgets across many channels.
Handling family members and multiple numbers
In India, a large share of enquiries come from relatives. Record the enquirer and the patient separately, allow more than one number per patient record, and ask at booking who the patient is. When matching to billing, match on the patient’s number and name, and keep the enquirer’s number as a linked contact. This single change often recovers a surprising share of unknown-source patients.
Other gaps worth closing, cheaply
- Phone number as the identity spine. Enforced and validated at registration in every unit, with the relative’s number recorded separately rather than instead. This single discipline fixes more measurement than any tool.
- Ask the source question properly. Not a twenty-option dropdown. Five options, asked out loud by the front office, with “saw something online” as one of them. Crude self-report beats a source field nobody fills.
- Dynamic number insertion on digital properties. Helpful but partial, because it only catches the patient who calls from the page rather than from memory.
- Accept lag deliberately. Report long-cycle service lines on a rolling quarter, never month on month, and say so in the review so nobody reads noise as trend.
- Reconcile to one source of truth monthly. Registrations from the hospital information system, not the CRM and not the platforms. Whatever cannot be reconciled to a registration is a claim, not a conversion.
What attribution will not tell you
- The value of brand and reputation, which shape choices long before any tracked touch.
- The influence of word of mouth and doctor reputation, beyond the referral you record.
- The effect of offline media, unless you use dedicated numbers or codes.
Use attribution to compare channels and fix leaks, and use separate measures such as awareness and preference for brand.
Holdouts: the only real test of whether spend caused volume
Attribution assigns credit. It cannot tell you whether a channel caused the patients it is credited with. The only way to know is to withhold spend somewhere and compare. In a multi-unit group the practical version is a geographic holdout: switch off one channel in one comparable catchment for a defined period, hold everything else constant, and compare total enquiries and registrations rather than platform conversions.
It is harder than it sounds, for reasons that have nothing to do with statistics. The unit head whose catchment is the holdout will escalate within a week. The agency will explain that the test is unsound; they are not always wrong, but the objection arrives suspiciously fast. Seasonality in Indian healthcare is strong and regional, so a four-week test spanning a festival is worthless. And in a group of a dozen units there may be exactly one valid pair of catchments.
I still think it is worth doing once a year on your largest channel. Brief the unit head personally and in advance, with the duration in writing and a fixed date to restore spend. The first time I ran one I skipped that step, and the test was cancelled in its second week by someone senior enough that I could not argue. A single hospital has no comparable catchment, so a time-based pause on one channel with a written restore date is the imperfect substitute.
Making budget decisions while attribution is uncertain
Stop trying to allocate precisely and start classifying. I split spend into three buckets and argue about the proportions rather than about line items.
| Bucket | What sits in it | How to judge it |
|---|---|---|
| Harvest | Brand search, your own listings, remarketing: anything intercepting demand that already exists | Cheap per conversion, low incrementality. Cap it rather than optimise it, or it will eat the budget. |
| Capture | Non-brand search, condition and procedure intent, aggregator presence, local discovery | Where attribution is least bad and most of the real work sits. Judge on cost per treated patient. |
| Create | Awareness in a new catchment, a new unit launch, a service line nobody knows you have | Aided awareness, branded search volume and total enquiry trend over quarters, never last-click. |
If you hold the create bucket to a cost-per-lead target you will zero it out within two budget cycles and then wonder why the new unit never matured. Decide the split as a judgement, document the reasoning and revisit it quarterly. A documented judgement the board can challenge is worth more than a precise-looking number nobody can interrogate.
If you are starting this next quarter
- Enforce phone-number identity at registration in every unit.
- Collapse the source dropdown to five options and train the front office to ask the question aloud.
- Reconcile one month of platform-claimed conversions against registrations and put the gap on a slide. That conversation changes how a group talks about attribution.
- Split the media budget into harvest, capture and create, and cap harvest explicitly.
- Move long-cycle reporting to rolling quarters.
- Plan one geographic holdout for the quarter after, with the unit head briefed in writing and a fixed restore date.
- Write down what you have decided not to measure and circulate it. Naming the unknowns is what stops someone inventing a number to fill them.
Attribution in healthcare is not a reporting problem you will solve. It is an uncertainty you manage, and the operators who do it well are the ones who stopped pretending otherwise early.
Further reading from primary sources
- Google Ads: About offline conversion imports: sending booked and treated outcomes back to Google Ads.
- Google Search Console: Performance report: measuring search queries, clicks and impressions.
Questions people ask
It links marketing touches such as searches, ads, WhatsApp chats and referrals to appointments, treatments and revenue, so channels are credited for treated patients rather than leads.
Lead attribution counts form fills and calls. Patient-journey attribution follows those enquiries into the hospital’s own records to see which became treated patients.
Start with source of enquiry linked to treatment. Move to multi-touch models only once enquiry and treatment data are clean.
A source on every enquiry, a CRM that captures calls and WhatsApp, a patient identity rule, and a link to appointment and billing data.
They credit conversions to their own clicks and count enquiries, not treated patients, and they cannot see the hospital’s billing data.
Use tracked numbers per channel or campaign and make sure every call creates or updates a CRM record with its source.
Record the referring doctor or source at the first contact and carry it through to the visit and billing records.
Yes. Personal data used for attribution needs a lawful basis and appropriate consent, and it must be protected like any other patient data.
Treated patients and revenue by source, cost per treated patient by channel, and the share of treated patients with an unknown source.
Yes. A consistent source list, a simple CRM and a monthly match against billing records are enough to start.
It shares credit across several touches in a patient’s journey rather than giving it all to the first or last one.
A high share of treated patients with no recorded source means the rest of the report should be treated with caution.
Usually the growth or digital team, with the CRM team maintaining the data and finance validating revenue.
It means switching off one channel in a comparable catchment for a set period, holding everything else constant, and comparing total enquiries and registrations rather than platform conversions. It is the only real test of whether spend caused volume. Run one a year on your largest channel, avoid festival periods, and brief the unit head in writing with a fixed restore date.
Classify rather than allocate precisely. Harvest (brand search, own listings, remarketing) is cheap but low in incrementality, so cap it. Capture (non-brand and procedure intent, aggregators, local discovery) is where attribution works best. Create (new catchments, units, service lines) is judged on awareness and enquiry trend over quarters, never last-click. Document the split and revisit it quarterly.
Hear the real question, which is usually whether spend is waste and how to defend it. A total-volume-against-total-spend trend by unit answers that better than any channel report. Give the CFO cost per identified enquiry and the enquiry-to-appointment rate instead, and do not use one deck for every audience.
Read my takes first in Google Search
