Making the digital and brand case for capex

The capital cases I saw in my first years inside a hospital group were built from two sources: the unit head’s occupancy sheet and a senior doctor’s conviction. “We are turning patients away.” “The cath lab is booked out.” “If we do not open in that city, someone else will.” Sometimes those assertions were right. The committee had no way of knowing, and neither did the people making them.

Meanwhile the digital function sat on the most direct evidence of demand the group possessed — what people were searching for, what they were asking the contact centre for, how long they were being told to wait, what they were saying about the competitor across the road — and none of it appeared in a single capital paper. Nobody had asked for it, and we had not thought to offer it.

This article is about closing that gap. Not because demand data replaces the occupancy sheet — it does not — but because a capital case with only supply-side evidence is asking a committee to fund a hypothesis. Demand evidence turns it into an argument. The trick is knowing how each kind of evidence is weighed by the people who sign, and presenting it in the form they can actually use.

What the committee is really deciding

A bed, a machine and a new market are three very different decisions that arrive on the same template. A bed extension is a capacity decision inside a demand pool you already serve. A machine — a linear accelerator, a robotic system, a second cath lab — is a mix and pricing decision: it changes what you can sell and to whom. A new market is a brand decision: it asks whether the name travels.

The committee is not asking “is there demand”. It is asking three narrower things. Is the demand we would serve incremental to the group, or moved from one unit to another? Will it arrive at the pace the payback assumes? And what happens to the number if the champion is wrong? Demand evidence from the digital front door answers the first two better than anything else in the building. It is almost silent on the third, which is why it cannot carry the case alone.

The five inputs the digital function actually owns

Each of these lives somewhere in your stack already. The work is extracting them cleanly and being honest about what each one can and cannot say.

Search demand

What people in a catchment are looking for, by specialty, procedure and location, over time. In an Indian city this means branded and generic terms across two or three languages, and it means separating “hospital near me” noise from procedure-specific intent. Its strength is that it is independent of your own capacity. A unit that is full will show flat enquiries because the contact centre has stopped taking them; search demand keeps rising regardless. Its weakness is that it is undifferentiated by payer, and finance knows that a large share of that intent is heading to a government scheme or a low-cost competitor.

Enquiry backlog

The enquiries the contact centre and the site received for a specialty or a unit that did not convert to an appointment, and the reasons logged. This is the closest thing you have to a queue of unmet demand. It is also the most contested, because the reasons field in a hospital CRM is, in my experience, filled in by whoever was on shift, using whichever option was nearest the top. Before you put a backlog number in a capital paper, get a sample re-listened to and re-coded. A number that says “no slot available” and holds up on audit is gold. One that does not survive the first challenge costs you the rest of the case.

Waiting-time data

Appointment lead time by doctor and department, and — if the hospital system exposes it — admission waiting lists and OT booking horizons. This is supply-side data but it lives on the digital side, because the booking engine sees it first. Rising lead time for a specialty, alongside rising search demand, is the pairing that makes a bed or machine case land. Lead time alone is ambiguous: it can mean demand, or it can mean one doctor who works two half-days.

Competitor listings

What the competitor set is advertising, listing and staffing, visible through their doctor pages, their aggregator presence, their maps listings and their paid search. This tells you what the market believes it can sell. A competitor who has just listed a new oncology unit and started bidding on the terms you are considering is evidence of two things: that they see the demand, and that the acquisition cost you are assuming is about to rise.

Review sentiment

What patients say about your unit, and the competitor’s, coded by theme. Complaints about waiting, about not getting a slot, about being sent to another unit — these corroborate a capacity case. Praise for a competitor’s new facility tells you what a market case is up against. Sentiment is the softest of the five and it should be used to colour a case, never to carry one.

How finance treats each one

This is the part I wish someone had told me earlier. Every one of those inputs gets discounted in a capital review, and the discount is different for each.

• Search demand is treated as market context. Interesting, not bankable. Finance will not let you convert search volume into revenue at any ratio you propose, because the ratio is yours and they cannot check it. Use it to establish direction and size of the pool, then stop.
• Enquiry backlog is treated as the nearest thing to a revenue line, and interrogated accordingly. Expect questions on leakage: how many of those enquiries went to a competitor, how many would have converted anyway at a lower price, how many are the same person calling three times. Have the answers before the meeting.
• Waiting time is trusted if it comes from the hospital information system and treated with suspicion if it comes from a dashboard the digital team built. Get the finance analyst to pull it themselves.
• Competitor listings are accepted as strategic context and used mainly to test your cost assumptions. They rarely change the decision. They do change the urgency.
• Review sentiment is heard and not weighed. Use it once, in a sentence, with a number of reviews behind it. Do not build a slide on it.

The pattern is simple. The closer an input sits to a booked appointment, the more it is worth. The further it sits from something finance can reproduce, the less.

The conversion ceiling is the argument

The most persuasive frame I have found for a bed or machine case is not “demand exists”. It is “the demand engine we already fund is hitting a ceiling, and here is the ceiling”.

Every group spends on acquisition. That spend produces enquiries. Enquiries convert at some rate into appointments, and appointments into admissions and procedures. When a specialty is capacity-constrained, the conversion rate drops for a reason that has nothing to do with marketing: the slot does not exist. You are paying to generate demand and then paying again in lost conversion to send it away.

Presented this way, the capital is not a bet on new demand. It is the cost of realising demand the group has already paid for. The finance team can check every step: acquisition spend by specialty, enquiry volume, conversion rate over time, the point at which conversion started falling while enquiries kept rising. The machine is the thing that lifts the ceiling. That is a case a CFO can defend to a board.

It also disciplines the ask. If conversion has not fallen — if the specialty is converting normally and simply wants more — then the capacity argument is weaker than the champion believes, and you should say so before someone else does.

Translating into the numbers they use

A capital paper runs on incremental cases, contribution per case, ramp and payback. Demand evidence has to arrive in those units or it stays in the appendix.

Incremental cases come from the backlog and the ceiling: the enquiries you are currently turning away, converted at the specialty’s historical rate, less a leakage allowance you have agreed with finance in advance. Contribution per case comes from finance, not from you; ask for it by payer mix and use their number. Ramp is where digital evidence adds something nobody else has — a new unit or machine does not fill on day one, and your own pre-launch and post-launch enquiry curves from previous openings are the best ramp assumption in the group. Payback is arithmetic once the others are honest.

Then the part most cases skip: cannibalisation. If the new bed, machine or unit draws from a catchment your existing units already serve, some of the “incremental” demand is a transfer. Search geography and enquiry pin-codes show this more clearly than any other source. Bring it up yourself. A case that shows its own cannibalisation and still works is far stronger than one that pretends there is none.

The market case is different

For a new city, the inputs change weight. Enquiry backlog is meaningless — you have no unit there. Waiting time is someone else’s. What matters is search demand by procedure in that catchment, competitor density and quality visible through listings and reviews, the share of that city’s patients already travelling to your existing units (which your CRM and admission data can show by pin-code), and whether the group’s name gets any branded search there at all.

That last one is the brand question, and it decides the demand plan. A name with branded search in a city it has never operated in can open with a shorter ramp. A name with none is starting from a doctor-led demand plan, and the capital case should carry a longer ramp and a larger pre-launch budget. I have watched a market case assume the flagship’s ramp in a city where nobody had heard of the group. It did not go well, and the digital evidence to prevent it was sitting in a search console.

What weakens the case

• Search volume presented as revenue at a conversion rate of your choosing.
• Backlog numbers that have never been audited against call recordings.
• Waiting-time charts from a digital dashboard rather than from the hospital system.
• A ramp curve borrowed from the best-performing unit rather than the most comparable one.
• Demand at its seasonal peak projected as the run rate.
• A single doctor’s enquiry volume presented as the specialty’s, when that doctor may leave.

The last one deserves a sentence of its own. In a doctor-led market, a lot of what looks like demand for a specialty is demand for a person. If the enquiries and the search terms name the doctor, the capital case is exposed to that doctor’s contract. Finance may not spot this. You will, and you should say it.

What you will be held to afterwards

The moment the digital function supplies demand evidence for a capital case, it inherits part of the outcome. That is the right trade, but go in with eyes open. Agree, before approval, which measure you will report against — enquiry volume for the specialty, conversion recovery, occupancy of the new capacity, cases on the machine — and at what checkpoints. Agree who pulls the number. Then report it every month, especially in the months when it is behind plan, with what you are doing about it.

The alternative is that the capital case becomes the unit head’s problem and the demand plan becomes yours, and when the two do not reconcile the conversation is about whose fault it is. Better to be in the same review from the start.

If you are building a capital case next quarter

1. Find out which capital decisions are in the queue for the coming cycle. Pick one where the demand story is genuinely in your data.
2. Pull search demand for the relevant specialty and catchment over at least two years, in every language that matters, and separate branded, generic and doctor-named intent.
3. Extract the unconverted enquiries for that specialty. Re-code a sample from recordings. Keep only what survives.
4. Ask the finance analyst to pull appointment lead time and OT horizon from the hospital system themselves.
5. Plot enquiries against conversion over time and find the ceiling. If there is no ceiling, tell the champion early.
6. Map enquiry and admission pin-codes against existing units to size cannibalisation, and put it in the paper.
7. Take the incremental-case number to finance for contribution and payback. Use their per-case figure, not yours.
8. Agree the post-approval measure, the checkpoints and who reports them, in writing, before the committee sits.

The doctor’s conviction still belongs in the room. It just should not be the only evidence that does.

Questions people ask