The specialty mix conversation, with the demand data on the table
Every hospital group has a specialty that gets more than its share. It has the newest equipment, the loudest advocate on the medical board, the most brochure space, and a place in every investor conversation. It is usually cardiac sciences or oncology or neurosciences, occasionally orthopaedics, and its position is defended with a mixture of clinical pride, historical revenue and the belief that it is what the brand stands for.
Every group also has a specialty that quietly fills beds, converts enquiries at twice the rate of the flagship, brings patients back for years, and gets nothing. Nobody argues for it because its head is not on the executive committee and its numbers are folded into “other”.
The demand data knows which is which. The problem is getting it onto the table in a room where the people who decide capital allocation have already decided what the hospital is for. This is about how the data reframes the mix conversation, why it is dangerous to bring, and how to bring it without starting a war you cannot win.
What the demand data actually says
The specialty mix in most groups is decided on three inputs: historical revenue by department, the medical board’s view of clinical priority, and what the competitor down the road just announced. Demand — what patients in the catchment are actually looking for, enquiring about and converting on — is not one of the inputs, because until recently nobody had it in a form that could be put next to a revenue line.
Now you do. The digital front door produces, for every specialty, a set of numbers that describe demand from the outside in:
- Search volume by specialty and condition in each city, and the group’s share of it — the proportion of those searches that reach a group page or listing.
- Enquiry volume by specialty through the site, the app, the contact centre and the aggregators, and where it is coming from geographically.
- Conversion from enquiry to appointment to admission, by specialty, which varies more than most people expect.
- Cost to acquire a patient by specialty through paid channels, which varies even more.
- Repeat and referral behaviour from the CRM — which specialties bring a patient back, and which bring their family.
Put these five next to the revenue and margin lines by department and the picture changes. The flagship often has high search volume that the group is capturing badly, low conversion because the enquiries are complex and the contact centre cannot handle them, a high acquisition cost, and a patient who does not return. The quiet specialty has moderate volume that the group captures well, high conversion, low acquisition cost and a patient who comes back twice a year and brings her parents. One of these is where the next unit of investment should go, and it is not the one with the brochure.
The data does not say the flagship is wrong. It says the assumptions underneath its position have never been tested, and that the specialty everyone ignores may be carrying more of the P&L than anyone credits.
Why nobody wants this conversation
Because the specialty mix is where clinical identity, personal standing and money meet, and the data threatens all three.
The head of the flagship built his department over a decade and has a reasonable claim that the brand exists because of it. He also sits on the medical board and has the CEO’s ear. When a growth leader arrives with a chart showing that his specialty converts poorly and costs three times as much to acquire a patient in, he hears an attack on his department dressed up as analytics. He will point out — correctly — that a cardiac patient is worth more than a paediatric one, that his cases are complex and enquiries take longer, and that the person presenting the chart has never treated anyone.
The medical director hears a marketing function trying to set clinical strategy. The CFO hears a new set of numbers that do not reconcile with the ones she already has and wonders which to trust. The unit heads hear that their capital requests are about to be re-prioritised by someone who does not run a hospital. And the CEO hears a fight starting that he did not ask for.
All of these reactions are rational. The data is being used to reframe a decision that several powerful people believe is already settled, and it is being brought by the function with the least clinical standing in the room. If you present it as a verdict, you will lose. If you present it as an input the room did not previously have, you have a chance.
The mistakes I made first
I brought the full deck the first time. Every specialty, ranked, with acquisition cost, conversion and share of search on one slide. It was accurate. It was also received as a league table, and the specialties at the bottom spent the next month attacking the methodology rather than the finding. One department head had his team pull the raw enquiry data and found — again, correctly — that some enquiries had been tagged to the wrong specialty by the contact centre. The error was small. The damage to the credibility of the whole exercise was not.
The second mistake was leading with acquisition cost. It is the number a growth leader finds most compelling and the one a clinician finds most offensive, because it turns his patients into a cost per unit. It is also the easiest to argue with, since it depends on channel mix and attribution assumptions that anyone can question.
The third was doing it in a room where the department heads were present and the CFO was not. The conversation needs the person who owns the capital allocation and the numbers that will be used to defend it. Without her, it is a debate between marketing and medicine, and medicine wins those.
How to present it without a war
Start with one specialty, and choose the one where the data flatters. Show that the quiet specialty is capturing demand efficiently, converting well and retaining patients — and that it has been under-invested against that performance. Nobody fights for the underdog to be ignored. The department head becomes an ally, and the room learns to read the five numbers on a case where the conclusion is comfortable.
Then show the same five numbers for the whole group, but not ranked. Grouped. Specialties where the group captures demand well and converts well. Specialties where there is demand in the catchment and the group is not capturing it. Specialties where the group is capturing demand and losing it at conversion. Specialties where demand is thin regardless. This is a diagnosis, not a league table, and each group has a different remedy — more investment, better search presence, a fix in the contact centre, or an honest conversation about whether the catchment supports the ambition.
The flagship usually lands in the third group: demand captured, lost at conversion. That is a finding the department head can live with, because the remedy is not “spend less on cardiac” but “the enquiries you generate are being lost before they reach you, and here is where”. It puts growth and the department on the same side of the problem.
Lead with share of demand and conversion. Bring acquisition cost only when the room is ready and the CFO has seen the method. Never lead with it.
Reconciling with the finance view
The CFO has a department-level P&L that allocates revenue and cost by the department that billed. The demand data describes patients by what they searched for and enquired about. These do not line up, and if you do not reconcile them before the meeting, the meeting will be about the discrepancy.
A patient who searched for chest pain, enquired about a cardiologist, was admitted under general medicine and had a procedure billed to cardiac sciences appears in three places. Sit with finance, agree the mapping from enquiry specialty to billing department, accept that it will be imperfect, and document the assumptions. Then present the demand data and the P&L on the same page with the reconciliation noted. When the department head challenges a number, the CFO defends the method because she helped build it. That is a different meeting.
What the data changes in practice
It rarely changes the flagship’s status. It changes the marginal decisions, which is where the money actually moves.
The next consultant hire goes to the specialty that is capturing demand it cannot serve, rather than to the flagship as a matter of course. The paid budget shifts from the specialty with the highest acquisition cost and lowest conversion to the one where each rupee produces an admission. The contact centre gets a specialist desk for the complex enquiries the flagship is losing. The new unit in a Tier 2 city leads its launch with the specialty the catchment is searching for, not the one the group leads with in the metro. The next capital request for equipment comes with a demand case attached, and the ones that cannot show one wait a cycle.
None of these are dramatic. Together, over two or three planning cycles, they move the mix — and the group finds it has grown a second and third strong specialty without ever having declared war on the first.
The politics you cannot avoid
Some of this will be resisted whatever you do, and you should know where.
The department head whose specialty shows thin demand will argue that demand follows supply — build the department and the patients will come. Sometimes true. Show the search data for the catchment, show what the competitor with an established department is capturing, and let the CEO judge whether the group can build a market or is buying into one.
The unit head whose capital request is deferred will argue that the demand data is a metro phenomenon and does not apply to her city. Sometimes true, too, particularly where patients search in a regional language and the tooling under-reads it. Bring the regional-language search and the contact centre call reasons for her city specifically. If the data is thin, say so, and agree how to improve it rather than pretending it is conclusive.
And the medical director will want to know that clinical priority still has a place. It does. The demand data is one of the inputs to the mix, not the only one. A group can decide to invest in a specialty the catchment is not yet searching for, as a strategic bet. It should just know that is what it is doing, and price the demand-building cost into the case.
If you’re starting this next quarter
- Build the five numbers by specialty — share of search, enquiry volume, conversion, acquisition cost, repeat behaviour — for one unit first, and check them with the contact centre for tagging errors before anyone else sees them.
- Reconcile with finance on the mapping from enquiry specialty to billing department, and document the assumptions.
- Find the flattering case — the under-invested specialty that performs — and present that first, with the department head in the room and on your side.
- Group, do not rank. Four categories with four remedies. Never a league table.
- Bring the flagship’s conversion leak as a problem you want to solve with the department, not as a verdict on the department.
- Attach a demand case to the next planning cycle as a standard input for every capital and hiring request, so that the data becomes routine rather than an event.
The specialty mix will still be decided by people with clinical standing and long memories. Your job is not to overrule them. It is to make sure that the next time they decide, the demand is on the table with everything else — and that when they choose against it, they know they did.
The data does not pick the winners. It just makes it harder to pretend the losers were chosen on evidence.
Questions people ask
Specialty mix is how a hospital allocates capital, consultants, equipment and marketing across service lines — cardiac, oncology, orthopaedics, paediatrics and the rest. In most groups it is decided on three inputs: historical revenue by department, the medical board’s clinical priorities and what the competitor down the road announced. Demand — what the catchment actually searches for, enquires about and converts on — has rarely been an input, because nobody had it in a form that sat beside a revenue line.
Five numbers by specialty. Search volume by condition in each city and the group’s share of it. Enquiry volume across site, app, contact centre and aggregators, with geography. Conversion from enquiry to appointment to admission. Cost to acquire a patient through paid channels. And repeat and referral behaviour from the CRM. Put those next to revenue and margin by department and the flagship often looks captured badly and lost at conversion, while a quiet specialty is carrying more than anyone credits.
Because the specialty mix is where clinical identity, personal standing and money meet. The flagship head built his department over a decade, sits on the medical board, and hears a chart on conversion and acquisition cost as an attack dressed up as analytics — presented by someone who has never treated a patient. The medical director hears marketing setting clinical strategy. Those reactions are rational. Present the data as a verdict and you lose; present it as an input the room never had and you have a chance.
Start with the flattering case: the under-invested specialty that captures demand efficiently, converts well and retains patients. Its department head becomes an ally and the room learns to read the five numbers on a comfortable conclusion. Then show the whole group grouped, not ranked — capturing and converting well; demand present but uncaptured; captured but lost at conversion; thin demand regardless. Each group has a remedy. A league table produces a month of attacks on methodology.
It is the number a growth leader finds most compelling and a clinician finds most offensive, because it turns his patients into a cost per unit. It is also the easiest to argue with, resting on channel mix and attribution assumptions anyone can question. Lead with share of demand and conversion. Bring acquisition cost only once the room has seen the method and the CFO has signed off on it. I learned this by leading with it.
Before the meeting, not in it. The P&L allocates by billing department; the demand data describes patients by what they searched and enquired about. A chest-pain search that became a general medicine admission with a cardiac procedure appears in three places. Sit with finance, agree the mapping from enquiry specialty to billing department, accept imperfection, document the assumptions. Then when a department head challenges a number, the CFO defends the method because she helped build it.
Rarely the flagship’s status. It changes the marginal decisions, which is where money moves. The next consultant hire goes to the specialty capturing demand it cannot serve. Paid budget shifts to where each rupee produces an admission. The contact centre gets a specialist desk for the complex enquiries the flagship is losing. The Tier 2 launch leads with what its catchment searches for. Capital requests arrive with a demand case, and the ones without one wait a cycle.
Two or three planning cycles, so two to three years before it is visible. Building and checking the five numbers for one unit takes a quarter, including the tagging clean-up with the contact centre. Reconciling with finance takes another. The first presentation should be one specialty, then the grouped view at the next cycle. None of the individual moves is dramatic. Together they grow a second and third strong specialty without declaring war on the first.
Mostly analyst time and tooling you probably already own: search data, CRM exports, contact centre tagging and the aggregator dashboards. The real cost is the tagging clean-up — enquiries logged to the wrong specialty by agents — and the reconciliation hours with finance. Budget it from the growth function, not the departments, so nobody feels their money bought a case against them. The return comes through better-directed hires and paid spend, not through a line item.
Yes, with a caveat the unit head will raise: patients there often search in a regional language and the tooling under-reads it, so metro data does not transfer. Bring regional-language search and the contact centre’s call reasons for that city specifically. Where the data is thin, say so and agree how to improve it rather than pretending it is conclusive. The Tier 2 launch that leads with the specialty its catchment is actually searching for is one of the clearest wins.
When the group decides it is a strategic bet, and knows that is what it is doing. The department head will argue demand follows supply — build it and patients come. Sometimes true. Show the catchment’s search data and what the competitor with an established department is capturing, and let the CEO judge whether the group can build a market or is buying into one. Then price the demand-building cost into the case rather than assuming it away.
The growth or digital function builds it, because the numbers come from the digital front door, the contact centre and the CRM. Finance co-owns the mapping to billing departments. The medical director keeps clinical priority as an explicit input alongside it. What must not happen is marketing presenting it alone to department heads without the CFO present — that becomes a debate between marketing and medicine, and medicine wins those.
Share of catchment demand captured and converted, by specialty, against the capital allocated to it. That single comparison shows whether the flagship’s position rests on evidence or on history, and whether the specialty folded into “other” is being under-invested against its performance. The board should not use it to pick winners. It should use it to ask whether the next capital and hiring cycle had a demand case attached, and what happened when it did not.
Three. I brought the full ranked deck — every specialty with acquisition cost, conversion and share of search — and it was received as a league table; one department found a small tagging error and the whole exercise lost credibility. I led with acquisition cost, the number clinicians find most offensive. And I did it with department heads present and the CFO absent. Now: one specialty first, grouped not ranked, finance in the room.
