Mental health digital-first: why the front door has to be different
The mental health service line was the first one where the funnel I had built for the rest of the group did active harm. The enquiry form asked for a name and a mobile number. The contact centre called back within the hour, as it was trained to, and said the name of the hospital and the word psychiatry to whoever answered. The appointment reminder went by text to a shared family phone. We had built a front door that a person in distress had to walk through in full view of everyone they were hiding from, and then we reported a low enquiry-to-appointment number and asked why.
Every other service line in a hospital group rewards speed, identification and follow-up. Mental health punishes all three unless they are redesigned. The person searching at two in the morning does not want to be found. They want to find out whether it is safe to come, what it will cost, and whether anyone will know. Very often the person searching is not the person who needs help, and the family member doing the searching is frightened of what a call might set in motion.
This is what a digital-first mental health line has to look like when the stigma is designed in rather than wished away — the intake, the tele model, the contact-centre script, the brand tone, the hard limits on the CRM, and how you measure a line where the main number is the people who did not turn up.
Why the front door has to be different
The search behaviour is unlike anything else in the hospital. Queries are about feelings and situations, not conditions: not sleeping, cannot stop worrying, a son who will not leave his room, a husband who drinks. The word psychiatrist appears late, if at all. Sessions are private-browsing, mobile, late at night. There is heavy reading and almost no clicking. When a form is eventually filled, it is often filled with a false name.
The searcher has two fears the digital front door has to answer before anything else: what will happen if I make contact, and who will know. A page that leads with a booking button and a doctor’s photograph answers neither. A page that explains, plainly, what a first conversation involves, that it can be by video, that no one will call the house, that the cost is this, and that the record is held separately — that page gets read to the end, and it is the one that produces the appointment three days later.
Anonymity in the funnel
The design principle is that identification is deferred as far as the process allows, and that every step asks for the minimum. A first contact should be possible with a first name and a way to be reached that the person chooses — a chat channel, an email, a time to call and a number that is safe to call. No OTP wall before the person has learned anything. No mandatory patient registration before a first tele-conversation.
Where the group’s systems require a registration number to book a consultation, and most do, the product decision is whether the registration happens after the first conversation rather than before it, and whether the front-end can hold a provisional record until then. This is a real fight with IT and with the registration desk, because it breaks the group-wide rule that every appointment is a registered patient. I have lost that fight once and won it once. When I lost, the line under-performed for a year and everyone blamed the campaigns.
The family-member searcher needs a separate path. A page and a contact option built for “I am worried about someone”, which explains what the family can and cannot do, what a consultation for the person would look like, and how to raise it with them. It converts slowly and it is the most-read page on the line. It is also the one that never gets built, because nobody in the marketing team can see the lead in it.
Tele-first, with a way back to the building
For most people the first contact should be a video or voice consultation, not a visit. It removes the waiting room, the reception desk that announces the department, and the journey home from a building that a neighbour might have seen them enter. The clinical team decides who needs to be seen in person and when; the product’s job is to make the tele step the default and the step-up to in-person a smooth one, in the same record, with the same clinician where possible.
The tele model also changes the economics. Psychologists and counsellors can serve every unit in the group and every Tier 2 city from wherever they are. Psychiatrists, who are scarcer, can be reserved for the cases that need them. A session is a lower ticket than most hospital consultations and the frequency is higher, so the line looks like a subscription business in the P&L rather than a procedure business, and the metrics should follow.
The non-negotiable is the crisis path. A tele-first line that cannot handle a person in immediate danger is a liability, not a service. The protocol — who is called, what the national helpline number is, how the clinician escalates, how the contact centre escalates — has to exist, be trained, and be tested before the line takes a single enquiry. I do not own that protocol. I own making sure the product does not go live without it.
The contact-centre script
The general contact-centre script is built to identify, qualify and book, fast. For mental health it has to be rebuilt from the first line. Agents do not say the hospital’s name until they have confirmed they are speaking to the person who made contact and that it is a safe moment. They do not ask what the problem is. They explain what happens next, ask how the person would like to be contacted, and offer the earliest slot without pressure. There is no upsell, no health-check offer, no follow-up call sequence if the person does not respond.
The agents handling the line should be a small, trained pool, not the general queue. Part of that is quality. Part of it is that these calls are hard on the people taking them, and a rotation and a supervisor who understands the line are a cost you plan for. When we routed mental health calls through the general queue to save headcount, the agents defaulted to the general script within a week, because scripts are what they had been trained on, and the numbers showed it.
Regional language matters more here than in almost any other line, because the vocabulary of distress is local, and because the words for mental illness in many Indian languages carry a stigma the English clinical terms do not. The contact centre and the content should use the words families use, and clinicians should help choose them.
Brand tone
The category is full of two kinds of failure: the smiling stock photograph with a slogan about breaking the stigma, and the fear-led performance ad that promises to fix a child’s exam anxiety in six sessions. Both destroy trust. The first is not believed. The second is noticed by the clinicians, who then refuse to be associated with the line, and by the advertising standards process, which is right to notice it.
The tone that works is quiet, specific and practical. What a first session is like. Who the clinicians are, as people, with their sub-specialties in plain language. What it costs. How privacy is protected, concretely — where the record is held, who can see it, what appears on the bill and the reminder. No promises about outcomes. No urgency. The group’s brand can carry a mental health line if it speaks in that register, and the line can raise the group’s brand in return, because a hospital that handles this well is trusted with everything else.
Under the advertising norms, the same rules as the rest of healthcare apply: no claims of cure, no superlatives, no testimonials that imply results. In this category I would go further than the rules and not run patient testimonials at all.
What the CRM must never do
This is the section I would put in front of every marketing team and every CRM vendor before the line launches, because the default behaviour of a hospital CRM is precisely the behaviour that will cause harm.
- No retargeting of anyone who visited mental health pages or made an enquiry, on any platform, ever.
- No audiences, lookalike or otherwise, built from mental health data. The consent basis does not exist and the reputational exposure is total.
- No cross-sell. A person who came for counselling does not receive a health-check offer or a cardiology camp invitation because they are now “in the base”.
- No automated reminder or follow-up that names the department, the clinician’s specialty or the hospital in a way that reveals the purpose of the appointment. The person chooses the channel and the wording.
- No recall sequence. If the person does not come back, the clinician decides whether to reach out, not the system.
- No visibility of mental health records to marketing users. Access is restricted to the clinical team and the trained contact-centre pool, with an audit log.
Every one of these breaks a default setting in every CRM I have worked with, and every one has to be verified after each system update, because defaults come back. Under the data-protection framework the consent for this data has to be specific, and it is not the consent that was collected at general registration. Building this correctly is the largest single piece of work in launching the line, and it is invisible to the board.
Measuring a line where no-show is the main metric
Enquiry volume is a weak signal, because so many enquiries are exploratory and many are made under a false name. Enquiry-to-appointment is misleading, because deferring identification means the funnel does not look like the others. The number that describes the health of the line is the no-show rate on the first appointment, followed closely by the rate at which people who attended a first session attend a second.
No-show tells you whether the front door has produced enough safety for someone to actually arrive. Second-session rate tells you whether the first experience — the tele setup, the clinician, the privacy, the follow-up that did not happen — was one they would repeat. Time from first contact to first appointment matters too; a person who reached out at two in the morning and is offered a slot in eleven days has been lost, and a scarce psychiatrist roster is the usual reason.
The P&L conversation has to be had honestly. The line is low-ticket and high-frequency, capacity is people rather than beds, and it will not carry the margin of a surgical specialty. The case I have made for it is on three grounds: it is a real and under-served demand pool in every city the group operates in; it is a brand line, in that a group seen to handle it well is trusted more broadly; and it is a catchment line, because the family that comes for this comes for other things later. The third is trackable by cohort if the CRM is built with the restrictions above and still permits aggregate analysis. I would not promise it in the first year and I would not present the line as a margin engine at all.
The order of operations
- Write the crisis protocol with the clinical lead, train the contact-centre pool on it, and test it. Nothing launches before this.
- Rebuild the intake to defer identification: first name, chosen channel, safe time to call. Fight for provisional records if registration blocks it.
- Build the “worried about someone” path as a first-class page and contact option.
- Configure the CRM restrictions above, document them, and add a check after every system update.
- Stand up the tele-first roster with psychologists and counsellors across units and a psychiatrist step-up path in the same record.
- Rewrite the contact-centre script and staff a small, supervised pool with rotation. Choose the regional-language vocabulary with clinicians.
- Set the brand tone in writing, with examples of what will not run, and put every creative through medical affairs.
- Report first-appointment no-show, second-session rate and contact-to-appointment time monthly. Present the line as access and brand, not margin.
The person you are trying to reach is trying not to be reached. Build for that, or build nothing.
Questions people ask
Because every other line rewards speed, identification and follow-up, and mental health punishes all three. A form asking for name and mobile, a callback within the hour that says the hospital’s name and the word psychiatry to whoever answers, and a reminder texted to a shared family phone — that is a front door a person in distress must walk through in full view of everyone they are hiding from. Then the line reports poor conversion and blames the campaigns.
In feelings and situations, not conditions: not sleeping, cannot stop worrying, a son who will not leave his room, a husband who drinks. The word psychiatrist appears late, if at all. Sessions are private-browsing, on mobile, late at night, with heavy reading and almost no clicking. Forms are often filled with a false name. Very often the searcher is a frightened family member, not the person who needs help, and they need a separate path.
It means identification is pushed as far down the process as the systems allow and every step asks for the minimum. First contact needs a first name and a way to be reached that the person chooses — chat, email, or a safe time and number to call. No OTP wall before the person has learned anything. No mandatory patient registration before a first tele-conversation. Where the hospital system demands a registration number, fight for a provisional record until after the first session.
Because a video or voice first consultation removes the waiting room, the reception desk that announces the department, and the journey home from a building a neighbour might have seen. Clinicians decide who needs to be seen in person; the product makes tele the default and the step-up smooth, in the same record. It also changes the economics — psychologists and counsellors can serve every unit and Tier 2 city from anywhere, with scarcer psychiatrists reserved for cases that need them.
The crisis protocol. A tele-first line that cannot handle a person in immediate danger is a liability, not a service. Who is called, the national helpline number, how the clinician escalates, how the contact centre escalates — it has to be written with the clinical lead, trained into the contact-centre pool, and tested. The growth function does not own that protocol. It owns making sure the product does not go live without it.
Rebuild it from the first line. Agents do not say the hospital’s name until they have confirmed they are speaking to the person who made contact and that it is a safe moment. They do not ask what the problem is. They explain what happens next, ask how the person wants to be contacted, and offer the earliest slot without pressure. No upsell, no health-check offer, no follow-up sequence. Use the regional-language vocabulary families use, chosen with clinicians.
No. Use a small, trained pool with rotation and a supervisor who understands the line. Part of that is quality; part is that these calls are hard on the people taking them, and that is a cost to plan for. When we routed mental health calls through the general queue to save headcount, agents defaulted to the general script within a week, because scripts are what they had been trained on, and the numbers showed it.
No retargeting of anyone who visited mental health pages or enquired, on any platform, ever. No lookalike or other audiences built from the data. No cross-sell — no health-check or cardiology camp offers because the person is now in the base. No automated reminder that names the department, specialty or hospital in a way that reveals the appointment’s purpose. No recall sequence; the clinician decides whether to reach out. No visibility of records to marketing users, with an audit log.
The consent for this data has to be specific, and it is not the consent collected at general registration. Every CRM restriction — no retargeting, no audiences, no cross-sell, restricted access — breaks a default setting in every CRM I have worked with, and defaults come back after system updates, so each one must be re-verified after every update. Building this correctly is the largest single piece of work in launching the line and it is invisible to the board.
Quiet, specific and practical. What a first session is like, who the clinicians are as people with sub-specialties in plain language, what it costs, and how privacy is protected concretely — where the record sits, what appears on the bill and the reminder. No outcome promises, no urgency. Avoid both the smiling stock photograph about breaking stigma and the fear-led ad promising to fix exam anxiety in six sessions. I would not run patient testimonials in this category at all.
The number that describes the line’s health is the no-show rate on the first appointment, followed by the rate at which people who attended a first session attend a second. No-show tells you whether the front door produced enough safety to actually arrive. Second-session rate tells you whether the first experience was one they would repeat. Time from first contact to first appointment matters too — a slot in eleven days loses someone who reached out at two in the morning.
Honestly. It is low-ticket and high-frequency, capacity is people rather than beds, and it looks like a subscription business rather than a procedure business. It will not carry the margin of a surgical specialty. The case rests on three grounds: a real, under-served demand pool in every city; a brand line, because a group that handles this well is trusted more broadly; and a catchment line, because the family comes for other things later. Do not present it as a margin engine.
Because very often the person searching is not the person who needs help, and that family member is frightened of what a call might set in motion. A page and a contact option built for worried-about-someone — what the family can and cannot do, what a consultation would look like, how to raise it — converts slowly and is the most-read page on the line. It is also the one that never gets built, because nobody in marketing can see the lead in it.
Write and test the crisis protocol first; nothing launches before it. Rebuild intake to defer identification. Build the worried-about-someone path as a first-class page. Configure and document the CRM restrictions with a post-update check. Stand up the tele-first roster with a psychiatrist step-up in the same record. Rewrite the script and staff a small supervised pool. Set the brand tone in writing and route every creative through medical affairs. Report no-show, second-session rate and contact-to-appointment time monthly.
