Oncology online: why cancer patients search differently
The first time I looked at the search data for an oncology service line properly, I stopped trusting the funnel I had built for everything else. The queries were not “oncologist near me”. They were the name of a drug, the name of a stage, a question about whether a number on a biopsy report was bad, and a question about cost, typed at one in the morning from a city three hundred kilometres from the hospital. Almost none of them were typed by the patient.
Every other service line I had run behaved like a consumer category with a medical accent. Oncology behaves like a family decision made under duress over several weeks, with the hospital as one of five tabs open on a nephew’s laptop. The enquiry arrives late, informed, and already holding a diagnosis from somewhere else. The digital front door built for a knee replacement is the wrong shape for it.
This is what I have learned about how cancer patients and their families search, and what a service line’s digital presence, contact centre and CRM have to do differently. It is also about the commercial argument, because oncology is the line where acquisition cost looks worst in month one and best in month twelve, and you will have to explain that to a CFO.
How the searching actually happens
A diagnosis is usually delivered in a smaller hospital, a diagnostic centre or a general physician’s clinic. The patient goes home. That evening someone in the family — a son, a daughter-in-law, a brother in another city — starts searching, and what they search is the words on the report. The histology term. The stage. The drug the local doctor mentioned. Then the questions: is this curable, what does treatment cost, which hospital is best for this in this state.
Three things follow. The session is long — weeks, not minutes — and the same family returns many times, on different devices, from different cities. The searcher is not the patient, so anything that assumes the reader has the symptoms is wrong. And the family already has a diagnosis, so content built to explain what a symptom might mean is useless. They are past that. They need to know what happens next and whether you are the place it should happen.
The queries also split by intent in ways a general keyword plan misses. Clinical understanding: what the stage means, what the treatment options are. Practical: cost, duration, whether chemotherapy needs admission, whether the hospital has a day-care unit. Comparative: this hospital versus that one, this city versus the metro. Financial: insurance, government schemes, whether the package is cashless. A service line that only answers the first group has built content for the wrong moment.
The family is the customer
The person who calls is rarely the person who will be treated. The contact centre has to be scripted for that. The agent asks for the patient’s name and date of birth and the caller does not know the date of birth. The agent asks about symptoms and the caller has a report, not symptoms. The agent offers an appointment on Tuesday and the caller needs to check with three people and an employer in another city first.
The CRM has to be built for that too. One patient, several contacts, several numbers, several languages, several cities. The decision-maker is often not the first caller. The son who found you may hand over to the daughter who is a nurse, who will hand over to the uncle who is paying. If your CRM holds one mobile number per lead, you lose the case the moment it changes hands, and it always changes hands.
I made this mistake in the first version of an oncology intake. The form asked for the patient’s details and the caller filled in their own, because it was the only way to complete it. Half the records were wrong at source and we did not notice for a quarter, because the calls still happened. What fixed it was not a smarter form. It was a form that asked, first, who you are calling for.
Second-opinion behaviour
Most oncology enquiries to a large group are second opinions in substance, whether or not they are labelled that way. The family has a diagnosis and a proposed plan and wants to know whether to trust it, and whether coming to you is worth the disruption. The front door has to start from the report, not from the symptom.
That means document upload is the primary action, not the booking button. Reports, scans, prescriptions, photographs of handwritten notes. A stated turnaround. A named clinical reviewer, or at minimum a named coordinator. A response that says what the next step would be and what it would cost, in plain language, before anyone is asked to travel.
The commercial temptation is to gate all of this behind a consultation fee and a visit. Some groups do and it works when their name is already the answer to the comparative query. For most, the second opinion is the product that earns the visit, and the conversion path is upload, response, tele-consultation, then travel. Build it that way and measure each step. Build it as a visit-first funnel and you will see enquiries and no admissions, and conclude the campaigns are not working when in fact the front door is.
Regional language is not translation
The searcher may search in English. The patient and the paying relative very often do not speak it comfortably, and the moment the conversation moves to treatment, cost and consent, it moves to Telugu, Tamil, Kannada, Bengali, Hindi or Marathi. If the contact centre cannot hold that conversation natively, the family goes to the hospital whose coordinator called back in their own language, even if your programme is stronger.
Regional-language content is different from translated content. A translated page about a cancer type reads like a textbook to a family in a district town. A page written for that family — about what treatment involves, how long, what it costs, where to stay, what the government schemes cover — in their language, ranks for the questions they actually type and gets read to the patient aloud. I have found that the practical pages outperform the clinical ones in every language, and that nobody in a marketing team wants to write them because they are not glamorous.
The other language surface is the appointment reminder, the estimate letter and the discharge summary. If those arrive in English to a family that speaks Odia, the brand promise of care stops at the front door.
The tumour board as a digital product
The clinical strength of a serious oncology programme is the multidisciplinary meeting where surgeons, medical oncologists, radiation oncologists, pathologists and radiologists agree a plan. Families rarely know it exists. The way it appears to them is that a doctor said something on Tuesday and a different doctor said something slightly different on Thursday.
Making the board visible is a product decision. The patient-facing version is simple: your case will be reviewed by the whole team on this day, here is what happens, and you will receive one written plan afterwards. That single written summary — in the family’s language, listing who reviewed, what was recommended, what the options are and what each would involve — is the most persuasive document an oncology service line produces. It answers the comparative query better than any campaign. It also travels: it gets forwarded to the uncle who is paying, and to the local doctor who made the referral, who then refers again.
I do not own the clinical content of that document and would not pretend to. What I own is that it exists as a standard artefact, that it is produced within a stated time, that it reaches the family through a channel they use, and that the CRM records that it was sent and what happened next. Most groups have the meeting. Few have the product.
The ethics of content in this category
Oncology marketing in India has a long history of doing damage, and the norms have tightened for good reason. The rules on treatment claims, the advertising standards code and the professional conduct regulations rule out survival claims, superlatives, cure language, and anything that trades on fear. That is the floor. The judgement calls sit above it.
Do not run lead-generation forms on pages that explain what a stage means. A family reading about a prognosis is not a lead. Do not retarget someone who read a page about advanced disease with an offer. Do not build lookalike audiences from oncology enquiries at all; the data is too sensitive and the consent basis under the data protection law does not stretch to it. Do not publish patient stories that describe outcomes, even with consent, in a way that implies typicality. Do publish honest cost ranges, because the absence of them sends families to forums where the numbers are invented.
Drug names deserve a specific policy. Families search them. Pages that explain what a drug is, how it is given and what it costs are useful and legitimate. Pages that promote a drug, or a hospital as the place to get it, are not. Medical affairs should sign every page in this category before it goes live, and the checklist should be written down, because the agency will not know where the line is and the line moves.
Measuring a line where the funnel takes months
Last-click attribution dies in oncology. The family touched you eleven times over five weeks across three cities before anyone booked anything. Report by cohort, not by campaign: enquiries opened in a month, and what happened to them over the following ninety days. Upload to response time. Response to tele-consultation. Tele-consultation to first visit. First visit to treatment start. Where the cases are coming from, by distance and by referring doctor.
Take that to the CFO as a pipeline, the way a sales function would, because that is what it is. Oncology enquiries convert slowly and the cost per enquiry looks high against a general OPD benchmark. The value per converted case, the length of the treatment relationship and the referral it generates back into the same district are what justify the line, and none of that is visible in a monthly lead report. I have had to argue this in a budget review more than once. The argument only holds if the cohort data exists, which means the CRM was built for cases rather than leads from the start.
The financial conversation is part of the funnel
Cost is the second most searched question after the diagnosis itself, and it is where most oncology journeys stall. A family that cannot see a credible number assumes the worst and either delays or goes to a public institution. A family that gets an estimate, a list of what insurance and the state or central schemes may cover, and a named counsellor to call, moves.
Financial counselling is therefore a product step with an owner and a turnaround, not a back-office function. The estimate should be a document with a version number, in the family’s language, sent through the channel they use, and logged in the CRM. The counsellor should be reachable without going through the general contact centre. The TPA empanelment status for the relevant packages should be visible on the site and in the estimate. Every one of those is a conversion lever that sits outside marketing’s traditional remit, and every one of them is where the journey breaks when growth does not own the whole path.
The order of operations
- Pull six months of search and enquiry data for the line and classify by intent: clinical, practical, comparative, financial. Notice which group has no content.
- Rebuild the intake around “who are you calling for” and document upload. Allow multiple contacts per case with language and city recorded.
- Write the contact-centre script for a relative holding a report, not a patient describing a symptom. Staff the languages of your catchment, not the languages of your city.
- Agree the tumour-board summary as a standard artefact with the clinical lead: format, turnaround, language, channel. Log it in the CRM.
- Write the practical pages — treatment duration, cost ranges, day-care, accommodation, schemes — in the top three regional languages of the catchment before writing another clinical page.
- Put a written content policy through medical affairs: no outcome claims, no lead forms on prognosis pages, no retargeting, no audience building from oncology data.
- Make financial counselling a tracked step with a named owner and a versioned estimate document.
- Report to the CFO by cohort over ninety days, with value per converted case and referral source, and stop reporting monthly lead cost for this line.
The family will find you either way. What they find decides whether they come.
Questions people ask
Not oncologist near me. They search the words on the report — the histology term, the stage, the drug the local doctor mentioned — then whether it is curable, what treatment costs, and which hospital is best for this in this state. The searcher is usually a son, daughter-in-law or brother in another city, not the patient. Sessions run for weeks across devices and cities, and the family already holds a diagnosis from somewhere else.
Because the person who calls is rarely the person who will be treated. The caller has a report, not symptoms, does not know the patient’s date of birth, and needs to check with three people and an employer before booking. The decision-maker changes hands — the son who found you hands over to the daughter who is a nurse, who hands to the uncle who is paying. A CRM holding one mobile number per lead loses the case the moment it changes hands.
Who you are calling for. In the first version of an oncology intake I built, the form asked for the patient’s details and the caller filled in their own, because it was the only way to complete it. Half the records were wrong at source and nobody noticed for a quarter because the calls still happened. The fix was not a smarter form but one that starts with who you are calling for, then allows multiple contacts per case with language and city recorded.
Yes. Most oncology enquiries to a large group are second opinions in substance: the family has a diagnosis and a proposed plan and wants to know whether to trust it. So the front door starts from the report — upload of scans, prescriptions and handwritten notes, a stated turnaround, a named reviewer or coordinator, and a plain-language response on next steps and cost before anyone travels. The path is upload, response, tele-consultation, then travel.
Because the searcher may search in English but the patient and the paying relative often do not speak it comfortably, and the moment the conversation moves to treatment, cost and consent, it moves to Telugu, Tamil, Kannada, Bengali, Hindi or Marathi. Translated clinical pages read like textbooks to a district-town family. Practical pages written in their language — duration, cost, where to stay, what schemes cover — rank for what they type and get read aloud to the patient.
A tumour board is the multidisciplinary meeting where surgeons, medical and radiation oncologists, pathologists and radiologists agree one plan. Families rarely know it exists; they see one doctor say something Tuesday and another something different Thursday. The product is a single written summary in the family’s language listing who reviewed, what was recommended and what each option involves, delivered within a stated time. It answers the comparative query better than any campaign and gets forwarded to the referring doctor.
Survival claims, superlatives, cure language and fear are ruled out by the treatment-claims rules, the advertising code and professional conduct regulations. Above that floor: no lead forms on pages explaining what a stage means, no retargeting anyone who read about advanced disease, no lookalike audiences from oncology enquiries because the consent basis does not stretch, no patient stories implying typical outcomes. Do publish honest cost ranges. Medical affairs should sign every page against a written checklist.
With a written policy, because families search drug names constantly. Pages that explain what a drug is, how it is given and what it costs are useful and legitimate. Pages that promote a drug, or promote the hospital as the place to get it, are not. Medical affairs should sign every page in this category before it goes live, and the checklist should be written down because the agency will not know where the line is, and the line moves.
By cohort, not by campaign. Last-click attribution dies when a family touches you eleven times over five weeks across three cities. Report enquiries opened in a month and what happened to them over ninety days: upload to response, response to tele-consultation, tele-consultation to first visit, first visit to treatment start, and where cases come from by distance and referring doctor. This only works if the CRM was built for cases rather than leads from the start.
Present it as a pipeline, the way a sales function would. Oncology enquiries convert slowly and cost per enquiry looks high against a general OPD benchmark, but value per converted case, the length of the treatment relationship and the referrals generated back into the same district are what justify the line. None of that appears in a monthly lead report. Stop reporting monthly lead cost for this line and report cohorts over ninety days with value per case and referral source.
Because cost is the second most searched question after the diagnosis, and it is where most journeys stall. A family that cannot see a credible number assumes the worst and delays or goes to a public institution. A family that gets a versioned estimate in their language, a list of what insurance and state or central schemes may cover, and a named counsellor reachable without the general contact centre, moves. Make it a tracked step with an owner and a turnaround.
A script written for a relative holding a report, not a patient describing a symptom. Agents should expect the caller not to know the patient’s date of birth, to have a histology term rather than symptoms, and to need time to consult family in another city. Staff the languages of your catchment, not the languages of your city, because the family goes to the hospital whose coordinator called back in their own language, even if your programme is stronger.
Pull six months of search and enquiry data and classify every query by intent: clinical understanding, practical, comparative and financial. Notice which group has no content — usually the practical and financial ones. Then rebuild intake around who you are calling for and document upload, rewrite the script for a relative with a report, agree the tumour-board summary as a standard artefact, and write practical pages in the top three regional languages before another clinical page.
