An Almaty clinic receives a hundred bookings from advertising, but its report shows only sixty visits. The owner proposes winning back forty "lost patients" with a message about another examination. The administrator opens the schedule: twenty appointments are still ahead, twelve patients cancelled and eight did not attend without cancelling.
All numbers in this example are illustrative. The marketer needs to investigate booking and attendance; the clinician and patient determine the need for an examination. Increasing LTV through unnecessary services would mix an operational task with a medical decision.
The page helps people reach an appropriate appointment
People need straightforward answers: who will see them, what is included, where the clinic is, what the appointment costs and when time is available. If the price covers only consultation, that is visible before booking. Additional services are not concealed behind a total.
The page's schedule must match available slots. The marketer and administrator follow the booking path and establish what happens when the chosen time is already taken. Promising tomorrow is unhelpful if the clinician is available only later.
The clinic provides preparation instructions approved by a medical professional. The marketer can make the text clearer and deliver it at the right moment, but does not change the advice to simplify a form or raise conversion.
The WHO approach to people-centred care centres people's needs and preferences, accessibility and coordinated care.
Separate dates and statuses first
Our group contains a hundred unique new patients, each with one initial booking. By the report date, eighty appointment times have passed. Sixty took place, twelve were cancelled by patients and eight ended in no-shows without cancellation. Twenty remain in the future. This example has no reschedules or clinic cancellations.
| Initial booking status | Patients | Place in the report |
|---|---|---|
| Appointment took place | 60 | Completed visit |
| Patient cancelled | 12 | Cancellation with an elapsed appointment date |
| No-show without cancellation | 8 | No-show with an elapsed appointment date |
| Date has not arrived | 20 | Future booking |
| Total | 100 | Entire original group |
So far, 60 / 100 = 60% of the group have attended. Among bookings whose dates have passed, attendance is 60 / 80 = 75%, keeping cancellations in the denominator. No-shows without cancellation among those same bookings are 8 / 80 = 10%.
Another calculation excludes cancellations: 60 / (80 - 12) = 60 / 68, about 88.2%. It measures attendance among uncancelled bookings whose dates have passed. A report cannot replace 75% with 88.2% while keeping the previous metric label.
In production records, rescheduling stays linked to the initial booking. Keep the old date, new date and reason. Record clinic-side cancellation separately too; otherwise a patient may be labelled absent from an appointment the service itself cancelled.
What bookings and completed visits cost
Suppose advertising for this group cost 300,000 ₸. Bookings are linked to the campaign under the agreed counting rule, and repeat enquiries from one patient do not create new people.
The cost per initial booking is 300,000 / 100 = 3,000 ₸. Ad spend per visit completed so far is 300,000 / 60 = 5,000 ₸. This is the current outcome for the whole group, with twenty appointments still ahead. Their outcomes may change the cost.
The 5,000 ₸ figure cannot be compared without explanation with a group whose appointment dates have all passed. Channel comparisons need the same observation period or data-readiness rule. Future visits stay visible until their outcome is known.
These amounts include only advertising. Administration, creative production and other acquisition costs enter the full calculation separately. To assess the economics of care delivered, the clinic adds actual receipts and the costs of delivering that service. A visit count cannot become profit without those records.
Repair begins with a reason
The twelve cancellations and eight no-shows may have different circumstances. Inconvenient times, an unclear price, difficulty finding the address and no way to reschedule call for different repairs. A status alone does not explain the reason.
The administrator records the operational reason the person gives, leaving unknown reasons unknown. The marketer looks for something to improve: time confirmation, an explanation of what the price includes or a clear rescheduling path.
Choose one change and observe comparable groups. Changing schedules, prices and reminders together cannot establish that the message caused the result. Account separately for clinician availability: more enquiries cannot help when suitable time slots are unavailable.
Reminders fit the clinic's process and patient preferences. Avoid sending a diagnosis or sensitive service name to a shared family phone for personalisation. A practical message can help with the date and address without those details.
Ads explain care without promising an outcome
"We solve every problem in one appointment" promises something marketing cannot control. Copy can accurately state the consultation format, professional qualifications, booking process and known terms. A relevant clinician reviews medical claims.
Fear and manufactured urgency should not replace explanation. An advertising page does not provide an individual diagnosis. An ad for a clinic licensed for medical practice states the licence number and the issuing authority (Art. 6, para. 4 of Kazakhstan's law On Advertising, as amended up to 17 May 2026; radio ads are exempt). The team checks the other requirements for advertising the chosen service, and any treatment claims, with the responsible specialists before launch.
Reviews describe one person's experience. A positive review cannot establish a typical clinical outcome. Public complaint responses do not disclose that someone was treated, a diagnosis or case details; the clinic offers a secure channel for investigation.
Support and marketing agree who responds and how to transfer the enquiry. People should not have to post medical details in a comment to get an operational problem noticed.
Separate booking data from advertising data
The attendance table needs aggregate counts by status and source. Diagnoses, examination results and complaint text do not belong in it. Information needed for care remains in its designated process with defined access.
Before analytics is connected, the developer and data owner inspect page addresses, event parameters and outgoing requests using test records. Free-text form content must not automatically enter a URL or advertising event. A field's absence from the visible form does not establish that an integration never sends it.
The Google Ads personalised advertising rules restrict targeting for sensitive categories, including health categories. For such offers, restrictions affect advertiser-built audiences, including Customer Match and first-party data segments. Before launch, establish the service's category and which promotion methods are permitted for it.
This is not a blanket ban on every clinic ad. Patient consent does not remove platform restrictions either. Every account has a list upload button, but if the service falls into the health category, Customer Match is not available for it.
A repeat visit follows a need
An appointment has two normal outcomes: a prescribed follow-up or completed care.
Records distinguish a prescribed follow-up, a new need and completed care. The clinician and patient decide whether another visit is needed. A revenue target cannot replace that decision.
Customer retention work cannot simply import a shop's "buys more often, therefore better" rule. No new visit after care is complete may be a normal outcome. Payment frequency does not measure treatment quality either.
The clinic can assess acquisition costs and contribution from delivered services by group. Keep booking accessibility, communication clarity, complaints and staff workload beside them. Medical outcomes need their own definitions and professional assessment beyond ad-account conversion counts.
Exercise: repair the forty "lost patients" report
Build an initial-bookings table for one group in your project. If you work outside healthcare, use another appointment-based service; the exercise needs no medical details.
Forty people in the original group have not yet attended, but only eight are no-shows without cancellation. Twenty appointments are still ahead and twelve were cancelled. Attendance among elapsed appointment dates is 75%, while current ad spend per completed visit is 5,000 ₸.
That report gives the team work on schedules, cancellations and reasons for absence. A bulk offer for another examination does not repair the accounting error. Choose the next useful step from the operational cause, leaving the medical decision with the clinician and patient.