Call tracking and offline conversion import for hospitals
Healthcare conversion tracking should tell Google Ads which clicks and calls became booked and honoured appointments, not just form fills. Capture click identifiers on landing pages, route calls through tracked numbers into the CRM, and upload status changes through offline conversion import or enhanced conversions for leads, while keeping clinical details out of every upload.
Google’s bidding will chase whatever you tell it counts as success. In most hospital accounts I review, that is a form fill or a call longer than some threshold, and the algorithm dutifully finds more of both, including the wrong number dialled, the job seeker and the patient who never turns up. Healthcare conversion tracking fixes this by telling the platform which clicks became booked and honoured appointments, using the hospital’s own CRM as the record.
This is the mechanics piece: calls, click identifiers, offline conversion import and enhanced conversions for leads, with the India-specific limits you need to plan around. The conceptual side, what attribution can and cannot tell a hospital, is in attribution in healthcare: what you can know. The wider paid programme sits in the complete guide to Google Ads for doctors in India.
What healthcare conversion tracking should actually count
Start with a ladder, not a single event. Every enquiry from an ad moves through stages, and each stage can be a conversion action in Google Ads:
- Lead: a form submission, a call of meaningful length, or a WhatsApp conversation started from the ad.
- Qualified: a real person with a genuine need your unit serves, confirmed by the desk or contact centre.
- Booked: an appointment with a date, doctor and unit in the scheduling system.
- Honoured: the patient arrived and the consultation happened, as recorded in the hospital information system.
Only one or two of these should be “primary” conversions that bidding optimises towards. The others stay “secondary” so you can see them in reports without the algorithm chasing them. Which rung you optimise to depends on volume: a busy multi-specialty account can often bid on booked appointments, while a single clinic may need to bid on qualified leads until the numbers build up.
The rung that matters most for the business is honoured appointments, which is why I prefer cost per honoured appointment as the headline efficiency metric. Getting that rung into Google Ads is what the rest of this article is about.
Call tracking in a hospital: the channel you cannot see by default
In many Indian specialties, a large share of paid enquiries arrive by phone, and phone calls are where tracking usually breaks. There are four ways to measure them, with very different quality.
Calls from ads
Call assets and call ads can use a Google forwarding number, a unique number from Google that tracks calls generated by your ads. India is on the list of supported countries. Google’s documentation also notes that in India, on some calls the caller ID shows a Google number rather than the caller’s own number, which matters if your desk calls people back from caller ID.
Website call conversions
A snippet on your site swaps your displayed number for a Google forwarding number when the visitor arrived from an ad, and counts calls above a minimum duration you set. Useful, but it only counts the call; it knows nothing about whether an appointment followed.
Clicks on a phone number
Counting taps on a mobile number is the weakest signal. It measures intent to call, not a call, and certainly not a booking. Keep it as a secondary action at most.
Imported call conversions and third-party call tracking
Google lets you import phone call conversions using the caller’s number and the call start time, so a call that later became a booking can be credited. The catch for India is explicit in the documentation: importing call conversions is supported for only a portion of calls in India. Plan on the assumption that some calls will not match.
That is why many hospital teams use a cloud telephony or call tracking provider that assigns numbers by source or session, records the click identifier where available, and writes each call into the CRM as a lead. The call then climbs the same ladder as a form, and its outcome goes back to Google through offline conversion import. Whatever vendor you use, check that the call recordings and caller data stay in systems the hospital controls and that the vendor’s data processing terms are signed.
WhatsApp enquiries need the same discipline
WhatsApp is now the default channel for many patients, and it is the easiest place to lose the thread between ad and appointment. A button that opens a chat with a generic message tells you nothing about which campaign sent the patient. The chat then lives on an agent’s screen, and the booking gets typed into the HIS with no record of where it came from.
The fix is to treat a chat like a form. When the patient taps the WhatsApp button on your landing page, the page should first store the click identifier and campaign details against a short reference, then open the chat with a pre-filled message carrying that reference. If you use the WhatsApp Business API through a CRM or messaging platform, the incoming conversation can create a lead record with the reference attached, and the lead climbs the ladder like any other.
Two cautions. Do not put campaign names that reveal a condition into the pre-filled message; the patient sees it and may forward it. And agree with the contact centre that bookings made in chat are logged in the CRM, not only in the HIS, or your WhatsApp numbers will always look worse than they are.
Offline conversions in Google Ads: how the import works
Every ad click carries a Google Click ID (GCLID). Offline conversion import lets you send Google a record that says, in effect, “this GCLID became a booked appointment at this time”, and Google credits the right campaign, ad group and keyword.
The flow in a hospital looks like this:
- The landing page captures the GCLID (and UTMs) into hidden fields when the visitor submits a form or starts a tracked call or chat.
- The CRM stores the GCLID on the lead record.
- When the desk marks the lead as booked, and later when the HIS confirms the visit, the CRM stamps each status with a timestamp.
- A scheduled job sends those status changes to Google Ads as conversions against the matching conversion actions.
Two rules from Google’s import guidelines shape the design. Conversions uploaded more than 90 days after the associated click are not imported, and imported statistics take roughly three hours to show in the account. So the job should run at least daily, and any appointment booked for a date far in the future needs its booking conversion sent when it is booked, not when it is honoured.
On the plumbing: Google now recommends the Data Manager route for new implementations, supports the Google Ads API and CRM connectors, and describes file uploads as legacy. Its help pages also note that from 15 June 2026 API uploads of offline conversions move to the Data Manager API. If your agency or CRM partner built an integration before that date, ask them to confirm it has been migrated and is still importing.
Enhanced conversions for leads
For teams starting fresh, Google’s own advice is to begin with enhanced conversions for leads rather than GCLID-only imports. The idea is simple. When someone submits a form, the Google tag sends a hashed version of the email or phone number they entered. When that lead later converts, you upload the same hashed identifier from your CRM, and Google matches the two, even if the GCLID was lost or the person switched devices.
Three practical points for hospitals:
- The window is shorter. Google’s guidelines say enhanced conversions for leads uploaded more than 63 days after the click will not be imported, so long booking lead times need thought.
- Send the GCLID as well whenever you have it. Google explicitly recommends including it to improve accuracy.
- You are sharing hashed personal data with an advertising platform. Your consent notice at the point of collection must say so in plain language, and your privacy team should sign off. I cover the privacy lens in privacy-safe remarketing in healthcare.
Keep health information out of the pipe
Conversion tracking needs to know that a conversion happened, when, and optionally its value. It does not need to know why the patient came. Name conversion actions by stage (“Consult booked”, “Consult honoured”), not by condition. Do not pass diagnoses, report types or symptoms into Google Ads fields, custom variables or URLs.
If you want to see results by specialty, you already can: the campaign and ad group structure tells you which specialty the click came from. There is no reporting gain from putting clinical detail into the conversion itself, and a real privacy cost if you do.
The same applies to conversion values. If you assign values so that bidding favours higher-value specialties, use standard values per stage and specialty agreed with finance, not individual bill amounts. Bills vary, arrive late and can reveal more about a patient’s treatment than an advertising platform should ever hold.
The CRM is the real conversion system
Offline conversion import is only as good as the statuses feeding it. In practice, the hard part is not the API; it is getting the desk to mark leads consistently and getting the HIS to confirm arrivals back into the CRM. If “booked” means different things at two units, your bidding learns two different things.
Write the status definitions down, agree them with unit heads and the contact centre, and audit a sample of leads each month against the appointment register. My notes on what a hospital CRM is actually for cover the wider case. The enquiry to appointment funnel calculator is a quick way to see where each unit’s ladder leaks before you start sending those stages to Google.
Setting up healthcare conversion tracking, step by step
- Agree the conversion ladder and status definitions with operations.
- Decide which stage is primary for bidding, by campaign, based on volume.
- Capture GCLID and UTMs on every landing page form, and configure the Google tag for enhanced conversions for leads.
- Choose the call route: Google forwarding numbers, a call tracking provider writing to the CRM, or both.
- Create conversion actions in Google Ads for each stage, named by stage, not condition.
- Build the scheduled upload from CRM to Google Ads through Data Manager, the API or a supported connector.
- Run a test week: book test appointments, confirm each appears in the CRM and then in Google Ads.
- Switch bidding to the new primary action only after the data has been flowing cleanly for a few weeks.
- Audit monthly: match rate, upload errors, and a sample of leads checked against the appointment register.
The landing page end of this chain, where the identifiers are first captured, is covered in landing pages for doctor consultation ads.
What this will and will not tell you
Done well, this setup tells you which campaigns, keywords and ads produce appointments that happen, and lets Google bid for more of them. It lets you compare specialties on something closer to business value than lead volume, which is also how you should read any cost per lead benchmark.
It will not capture the patient who saw your ad, then walked in a week later, or the one referred by a family doctor after reading your page. Some calls in India will not match. Those gaps are why I treat platform-reported conversions as one input in patient-journey attribution, not the whole answer. A tracking setup that is honest about its blind spots is more useful to a CFO than one that claims to see everything.
Questions people ask
It is the setup that tells Google Ads which ad clicks and calls became real outcomes, ideally booked and honoured appointments rather than just form fills. It combines on-site tracking, call tracking and uploads from the hospital CRM, so bidding optimises towards patients who actually arrive. Done properly, it also gives leadership a cleaner view of cost per appointment by specialty and unit.
Because without it, the ad platform optimises towards cheap enquiries that may never become revenue. Importing booked and honoured appointments lets bidding favour the campaigns that produce visits, and lets finance compare spend with outcomes recorded in the hospital’s own systems. The cost is mostly integration and process work, not media, and it improves how every future rupee is spent.
The technical integration can be quick if the CRM already stores click identifiers. The slower part is operational: agreeing status definitions, training the desk to update leads consistently and connecting arrival data from the HIS. Allow a test period of a few weeks with real bookings before you switch bidding to the imported conversions, so the algorithm learns from clean data.
Hidden form fields to capture GCLID and UTMs, the Google tag configured for enhanced conversions for leads, a CRM field for click identifiers, a reliable feed of appointment and arrival statuses from the HIS, and a scheduled job or connector to upload conversions. IT should also confirm where data is stored, who can access it and that uploads carry no clinical information.
Yes, Google lists India among the countries where forwarding numbers are available for call reporting. Two caveats matter. Google notes that in India some calls show a Google number as caller ID rather than the patient’s number, and importing call conversions is supported for only a portion of calls in India. Many hospitals therefore also use a call tracking provider that writes calls into the CRM.
Google recommends it for new setups because matching on hashed email or phone numbers can credit conversions even when the click identifier is lost or the patient changes devices. It has a shorter upload window than GCLID import, so plan for long booking lead times. Most teams do best sending both: the hashed identifiers and the GCLID when available.
That the consent notice at the point of collection clearly explains that contact details may be used to measure advertising, that the purpose is specific, and that no health information travels in the upload. They should also review the vendor and platform data terms and confirm how a withdrawal of consent stops future uploads. This is an operational view, not legal advice.
The deepest stage that has enough volume for the algorithm to learn from. Large multi-specialty accounts can often optimise to booked appointments. Smaller clinics may need to optimise to qualified leads while booked and honoured appointments are tracked as secondary conversions. Revisit the choice as volume grows, and never optimise to raw clicks on a phone number.
Directly. The desk or contact centre marks leads as qualified and booked, and those statuses become the conversions Google learns from. Inconsistent marking teaches the algorithm the wrong things. Give agents clear status definitions, make the CRM screen quick to update during a call, and audit a sample of leads against the appointment register each month.
Values help when you want bidding to favour higher-value specialties, but they must be consistent and agreed with finance. Many hospitals use a standard value per stage and specialty rather than actual bill amounts, which vary and can lag by weeks. Avoid values that reveal clinical detail, and do not switch to value-based bidding until the underlying data is reliable.
The number of conversions uploaded by stage, upload errors and their causes, the match rate for imported calls and leads, any change in conversion action settings, and a reconciliation of reported bookings against the CRM. Ask for a short note on anything that broke and what was fixed. Tracking health deserves the same attention as spend and results.
Counting taps on a phone number as conversions, counting short or repeat calls, losing the link between a tracked number and the lead record, and leaving call assets live when the desk is closed. Another is using caller ID for callbacks without realising some calls in India may show a Google number. Each of these quietly distorts what bidding optimises towards.
No. It shows what can be tied back to an ad click or tracked call. Walk-ins after seeing an ad, referrals from family doctors and conversations that move to personal phones will be missing or partial. Treat platform conversions as one reliable input and combine them with CRM, HIS and survey data when judging the overall contribution of paid media.
Use one set of status definitions, one CRM configuration and one set of conversion actions across units, with the unit captured as a field rather than in separate setups. Standardise landing page forms and call routing so identifiers are captured the same way everywhere. Then compare units on the same stages, and investigate any unit whose match rate or booking rate drifts.
Read my takes first in Google Search
