The three numbers a board should ask growth for
The growth section of a hospital group’s quarterly board pack tends to arrive in one of two states. Either it is forty slides of channel metrics that nobody on the board can connect to the P&L, or it is a single slide of revenue growth that tells the board nothing about whether the engine behind it is healthy. Both are failures of the same kind. Growth has not decided what it is accountable for, so it is reporting everything, or reporting the outcome and hiding the mechanism.
I have produced both versions. The forty-slide pack was a defence — if every number was there, nobody could accuse me of hiding one. The one-slide pack was a different defence — if only the outcome was there, nobody could ask about the parts. Neither did the board any good. A board does not need to see the engine. It needs to know whether it is running, whether it is getting more expensive, and whether what it produces is worth having.
That is three questions. It should be three numbers. Here is how I chose mine, why they are hard to defend, what each one hides, and what it takes to stop the pack growing back.
Why three
A board meets four times a year for a few hours and has a dozen agenda items. The growth section gets perhaps twenty minutes. In twenty minutes, non-executive directors can absorb three numbers, compare them with last quarter, and ask one good question about each. They cannot absorb a funnel. They should not be asked to.
Three also forces a choice. If you can only show three, you have to decide what growth is for. In a hospital group the honest answer is: bring in patients the group would not otherwise have had, at a cost the margin can carry, and make sure they come back. Everything else — traffic, followers, rankings, containment rates, campaign reach — is a means, and means do not belong in front of a board.
The first number: cost to acquire a new patient
Fully loaded. Not cost per lead, not cost per click, not the media number. Take everything growth spends — media, agency, contact centre, CRM, listings, the digital team’s own cost — and divide it by the number of first-time patients who kept an appointment in the quarter and whose first touch was a channel growth owns.
This number is unpopular inside the function because it is much larger than the media-only figure and much harder to move. It is exactly the number the board should see, because it is the only one that can be set against the contribution margin of a new patient. When the cost to acquire crosses a threshold of what a new patient contributes in their first year, the group is buying growth it cannot afford, and the board should know before the CFO tells them.
Defending it means being able to say what is in the denominator. A patient who came through a doctor’s own reputation, a corporate empanelment, or a walk-in is not growth’s to count. The temptation to include them is enormous, because it halves the number. Resist it once, in front of the board, and you will never be able to un-resist it, which is the point.
The second number: demand the group could not serve
Every enquiry that arrived and did not become a kept appointment, expressed as a share of all enquiries, and split into two causes. The ones growth lost — unanswered calls, slow callbacks, a booking journey that broke. And the ones the hospital lost — no slot within a week, the doctor was not available, the department did not call back.
This is the number that makes a growth leader useful to a board rather than just to a marketing budget. It shows the conversion ceiling. It tells the directors that the engine is producing more demand than the units are converting, and it puts the reason on the table. In every group I have worked in, the second cause was larger than the first, and the board had never seen it because nobody in growth wanted to report a number that was mostly someone else’s problem.
The defence is data quality. You need the contact centre and CRM to record why an enquiry did not convert, consistently, across units, in every regional language the centre operates in. That is a year of unglamorous work before the number is trustworthy. Report it anyway, with the caveat, from the first quarter. A board that has seen an imperfect version of this number will fund the work to make it accurate. A board that has never seen it will not fund anything.
The third number: what a new patient is worth after twelve months
Take the patients acquired through growth channels four quarters ago. What have they spent with the group since? Not what they spent on the first visit — what the cohort has generated across every unit, every department, every format, in a year.
This is the number that connects growth to the P&L in the way the board actually thinks about it. It tells them whether the patients the engine brings in are the ones the group wants: the ones who stay, who bring their families, who move from an outpatient consult to a procedure. It exposes the difference between a channel that produces cheap first visits and a channel that produces patients.
It is also the number most groups cannot produce, because the patient identity is split across the hospital information system, the CRM, the app and three unit-level databases that were never reconciled. If you cannot report it, say so, and say what it would take. That single admission has done more to get data work funded than any AI roadmap I have presented.
What each number hides
Three numbers are a compression, and every compression loses something. A board should know what it is not seeing.
The acquisition cost hides mix. A falling number may mean the engine is more efficient, or it may mean growth has shifted spend toward a cheap low-acuity service line and away from the expensive high-value one. Report the blended figure but be ready to show the split by service line when asked, and make sure someone asks.
The unserved-demand number hides geography and language. A group-level figure can look acceptable while a Tier 2 unit is losing half of its enquiries because the contact centre has two speakers of the local language on a shift of thirty. The board does not need the unit split every quarter. It needs to know it exists and that someone is looking at it.
The twelve-month value hides time. It is a lagging number by construction. What it tells you about the engine is a year old. Pair it, when you can, with a leading indicator — repeat visit rate at ninety days, say — but do not put that on the board page. That is a management number.
All three hide the brand. None of them tells the board whether the group’s promise is being kept or eroded. There is no clean quarterly number for that, and I have stopped pretending there is. What I do instead is bring one qualitative page — what patients said this quarter, where the reviews turned, what the search data says people are asking — and I keep it to a page.
Defending the numbers when they move the wrong way
Choosing three numbers is easy in a good quarter. The discipline is tested when one of them deteriorates and you are standing in front of the board.
The wrong response is to add slides. A rising acquisition cost surrounded by twelve slides of channel detail reads as exactly what it is: an executive explaining why it is not his fault. The board notices, and the pack grows.
The right response is to say what moved it, what you have decided to do, and what the number will be in two quarters if the decision works. One sentence each. If the cause is outside growth’s control — a competitor opened across the road, an aggregator changed its terms, a unit lost its lead surgeon — say so plainly, and say what growth is doing about the part it does control. Boards forgive a bad number. They do not forgive an executive who cannot explain one.
I got this wrong for a year. Every deterioration became a case for the defence, and the case grew longer each quarter. What finally changed it was a chair who said, in a break, that he had stopped reading past page two. That is the sentence you should hear in your head when you add a slide.
How the pack grows back
It grows back one reasonable request at a time. A director asks about the app, and next quarter the app has a slide. The CFO asks about agency spend, and next quarter there is a vendor breakdown. The medical director wants to see the department-level numbers, and next quarter there are fourteen departments. Each addition is defensible. Together they are forty slides again within a year.
The only defence I have found is structural. The three numbers are the board page, agreed with the chair and minuted. Everything else lives in an appendix that is circulated but not presented, and a management pack that goes to the executive committee, where the detail belongs. When a director asks a question the board page cannot answer, the answer goes to the appendix, not the page. When the same question is asked three quarters running, that is a signal the three numbers might be wrong, and it is worth a conversation with the chair — not a fourth number.
Once a year, revisit the three. Growth’s job changes as the group changes. A group in launch mode cares about acquisition cost above all. A group with mature units cares about twelve-month value and unserved demand. The numbers should follow the strategy, not the reporting habit.
If you are rebuilding the pack next quarter
- Write down what growth is accountable for in one sentence. If you cannot, the pack problem is a mandate problem, and no amount of slide-cutting will fix it.
- Draft the three numbers and take them to the CFO before the chair. If the CFO cannot connect each one to a P&L line, change the number.
- Take them to the chair. Agree the page, agree the appendix, agree that the page does not grow without a conversation.
- Audit the data behind each number. Where it is unreliable, report it anyway with the caveat, and put the fix on your own roadmap with a date.
- Present the three for two quarters with the old pack in the appendix. In the third, drop the appendix to ten pages. Nobody will notice.
- When a number moves the wrong way, rehearse the three sentences — what moved it, what you decided, what it will be — before the meeting. Do not rehearse a defence.
A board that can hold three numbers in its head will hold you to them. That is not a risk to manage. It is the job.
Questions people ask
Three. The fully loaded cost to acquire a new patient. The share of demand the group could not serve, split between what growth lost and what the hospital lost. And what a patient acquired through growth channels is worth after twelve months, across every unit and format. Everything else — traffic, rankings, containment rates, campaign reach — is a means, and means do not belong in front of a board that has twenty minutes.
Take everything growth spends — media, agency, contact centre, CRM, listings, the digital team’s own cost — and divide by first-time patients who kept an appointment that quarter and whose first touch was a channel growth owns. It is much larger than the media-only figure and harder to move, which is why it belongs on the board page: it is the only number that can be set against a new patient’s first-year contribution. Never count doctor-reputation, empanelment or walk-in patients.
Every enquiry that did not become a kept appointment, as a share of all enquiries, split by cause. Growth’s losses: unanswered calls, slow callbacks, a broken booking journey. The hospital’s losses: no slot within a week, doctor unavailable, department never called back. It shows the board the conversion ceiling — the engine producing more demand than the units convert. In every group I have worked in the second cause was larger, and nobody in growth wanted to report it.
Because first-visit revenue tells you what a channel produced; twelve-month value tells you whether it produced patients. Take the cohort acquired four quarters ago and add what they have spent across every unit, department and format since. That exposes the difference between cheap first visits and patients who stay, bring family and move from consult to procedure. It is the number that connects growth to the P&L the way a board thinks about it.
Acquisition cost hides mix — a falling number may mean spend shifted to a cheap low-acuity line. Unserved demand hides geography and language; a group figure can look fine while a Tier 2 unit loses half its enquiries for want of a local-language agent. Twelve-month value hides time; it is a year old by construction. And all three hide the brand, which is why I bring one qualitative page on what patients said this quarter and keep it to a page.
About a year for unserved demand, because the contact centre and CRM have to record why every enquiry failed, consistently, across units and regional languages. Twelve-month value takes longer if patient identity is split across the HIS, CRM, app and unit databases. Report all three anyway from the first quarter, with the caveat stated. A board that has seen an imperfect version will fund the fix. A board that has never seen it funds nothing.
A reconciled patient identity across the hospital information system, the CRM, the app and whatever unit-level databases grew up separately, so a cohort acquired in one quarter can be followed across every unit and format for a year. Most groups cannot do this today. Say so to the board and say what it would take. That single admission has done more to get data work funded than any AI roadmap I have presented.
Three sentences: what moved it, what you have decided to do, and what the number will be in two quarters if the decision works. If the cause is outside growth’s control — a competitor across the road, an aggregator changing terms, a unit losing its lead surgeon — say so plainly and say what you are doing about the part you control. Boards forgive a bad number. They do not forgive an executive who cannot explain one. Do not add slides.
One reasonable request at a time. A director asks about the app; next quarter it has a slide. The CFO asks about agency spend; a vendor breakdown appears. Each addition is defensible and together they are forty slides within a year. The only defence is structural: the three numbers are the board page, agreed with the chair and minuted. Questions go to a circulated appendix and the executive committee pack. The same question three quarters running is a conversation with the chair, not a fourth number.
Growth reports it; the units own most of the remedy. That is exactly why it is worth reporting. The split by cause means the medical director and unit heads see that slots, doctor availability and departmental callbacks are the larger leak, without growth having to argue it. Take the number to the CFO before the chair — if she cannot connect it to a P&L line, change it — and take it to unit heads before the board sees it.
Not as a number, because there is no clean quarterly one and I have stopped pretending otherwise. The three numbers say nothing about whether the group’s promise is being kept or eroded. What I bring instead is one qualitative page — what patients said this quarter, where the reviews turned, what the search data says people are asking. One page. Leading indicators like repeat visits at ninety days stay in the management pack.
Yes, and they are easier to produce. A single site has one HIS and usually one CRM, so twelve-month value is within reach much sooner, and unserved demand needs only one contact centre to record reasons. The discipline is the same: agree the page with the chair, keep detail in an appendix, and revisit annually. A hospital in launch mode weights acquisition cost; a mature one weights unserved demand and twelve-month value.
Once a year, deliberately, and not because a director asked a question. Growth’s job changes as the group changes: a group in launch mode cares about acquisition cost above all; a group with mature units cares about twelve-month value and unserved demand. The signal that a number is wrong is the same question asked three quarters running. That is a conversation with the chair about replacing one — never a reason to add a fourth.
