Referral rate by source rather than referrer satisfaction. Leaving dates against known events rather than exit interviews. A source going quiet triggers a phone call. The people who left are not answering your survey, which is the one thing you can be certain of.
Our own project · Meridian Hospital Group is a constructed operator; the instrument and its figures are real
Meridian Hospital Group · Stakeholders
Nobody tells you they have gone
Every project produces a stakeholder map, and almost none of them are any use, because a list of interested parties says nothing about which of them can actually do something to you. Sorted by what each one can do, the map rearranges sharply, and the group that turns out to matter most is the one that never complains, never negotiates and never appears in a consultation: the people who can simply stop coming.
The entity page sets out an actor model with two actors, the patient and the provider, the second of which is a scale rather than a party. That model is deliberately narrow because it is doing a specific job: explaining why the cost of a decision is invisible to the person taking it.
It leaves out everybody else, and everybody else is quite a long list. This page is that list, sorted not by how important they sound but by what each one is actually able to do when they are unhappy.
Chapter 1 · Sort by instrument, not by importance
A list of interested parties is not an analysis
The standard stakeholder map arranges people by influence and interest, which produces four quadrants, a slide, and almost no decisions. The trouble is that influence is treated as a single quantity, and it is not. It is several different instruments that behave in completely different ways.
There are three, and the distinction is old enough to be reliable. Some people can leave. Some people can speak. Some people can do neither, and a proportion of those will eventually obstruct, because obstruction is what remains when the other two are unavailable.
Sorted that way the map rearranges, and the rearrangement is the useful part. The groups that consume the most management attention are the ones with voice, because voice arrives as words in a meeting and words demand a reply. The groups that do the most damage are the ones with exit and no voice, because they never say anything at all.
And the groups with neither are the ones an organisation will discover it has an obligation to only when somebody outside it asks a question, which for a hospital in this market is a matter of when rather than whether.
Chapter 2 · Twelve, by what they can do
The register
Twelve groups, each tagged with the instrument available to it. Read the third field, which is how you find out you have lost them, because that field is the whole argument of this page.
Notice how many of them read some version of months later, with no reason attached. Those are the relationships that will be managed badly, not through neglect but because there is no moment at which management is triggered.
-
The people who put up the capital
Exit and voice- What they want
- A return, at an acceptable risk, on a horizon considerably shorter than the life of the building.
- What they can do about it
- Both. They sit on the board and they can sell. The combination makes them the only group whose interests are automatically represented in every decision without anybody having to arrange it.
- How you find out you have lost them
- Immediately and in words. This is the one relationship where nothing is ever discovered late.
-
Referring clinicians outside the hospital
Exit only, and silent- What they want
- Their patient looked after, a report back, a colleague who returns a call, and not to be embarrassed by having sent somebody somewhere bad.
- What they can do about it
- Nothing except stop. They have no contract, no vote and no complaint process, and they are under no obligation to explain. The most trusted voice in the whole decision simply points somewhere else.
- How you find out you have lost them
- A referral rate that fell eighteen months ago, noticed as a budget variance, explained as market conditions.
-
Patients who can pay
Exit only- What they want
- To be seen quickly, treated competently, not harmed, and to understand what is happening. Failing that, a flight.
- What they can do about it
- Leave, and in this market leave the country. They are the demand case and they are under no obligation to tell anybody why they went.
- How you find out you have lost them
- A market share number, months later, with no reason attached to it.
-
Patients who cannot pay
Neither- What they want
- The same things, and they are ill at the same rate as everybody else.
- What they can do about it
- Nothing at all. They are not customers, so exit is meaningless, and they have no route to voice because they were never in the relationship.
- How you find out you have lost them
- You do not. They are absent from every figure the hospital produces, which is the definition of the third kind of cost.
-
Senior clinicians
Exit, and weak voice- What they want
- A department that works, equipment that is present, peers, a career that goes somewhere, and to be paid properly.
- What they can do about it
- Accept the counter offer. Their voice inside the organisation is real but costly to use, and their exit is cheap, fast and final.
- How you find out you have lost them
- A resignation, which arrives after the decision to leave was taken, usually months after the thing that caused it.
-
Nurses and support staff
Exit, and voice that is expensive to use- What they want
- To do the job to the standard they were trained to, to finish near enough on time, to be safe, and not to carry the gap between the care they could give and the care they gave.
- What they can do about it
- Leave, which they do, and which arrives at the employer as agency spend under a heading that describes the symptom. Raising a concern costs them something; leaving costs them less.
- How you find out you have lost them
- Turnover, sickness and agency rates, all of which are lagging and all of which are filed as workforce problems rather than as the design or staffing decisions that produced them.
-
The regulator
Voice only, and it is absolute- What they want
- Compliance with a standard, evidence of it, and no surprises.
- What they can do about it
- Nothing except speak, and everything, because the licence is a binary. They cannot leave and they do not need to.
- How you find out you have lost them
- In writing, with a deadline. The most legible stakeholder on this list.
-
Professional bodies and unions
Voice, and collective exit- What they want
- Terms, conditions, safety and standing for their members, and a say in decisions that affect all three.
- What they can do about it
- Negotiate, and withdraw labour collectively. In this market that is a live and recurring feature rather than a remote possibility, and it converts a staffing disagreement into a total shutdown.
- How you find out you have lost them
- Loudly and in advance, which is a feature rather than a nuisance. A stakeholder who tells you before acting is doing you a service.
-
The host community
Obstruction- What they want
- Employment, respect, a share of what arrives, and not to have their road blocked by ambulances and their nights filled with sirens.
- What they can do about it
- Neither leave nor be heard through any formal channel, so the available instrument is disruption: the blocked gate, the stalled delivery, the site that cannot be worked. Obstruction is what people use when the other two are missing.
- How you find out you have lost them
- All at once, on a day you did not choose, usually over something small that stood for something large.
-
Suppliers, and the accredited distributor
Exit, which under a single source policy is leverage- What they want
- Volume, payment on time, and a relationship worth protecting.
- What they can do about it
- Stop supplying, or reprice. A sourcing policy that refuses the spot market deliberately hands the accredited supplier a position, and that is the cost of the control rather than an argument against it.
- How you find out you have lost them
- At renewal, in the price, or at the worst possible moment in a shortage.
-
The partner institution
Exit- What they want
- Placements, teaching capacity, research, and a partner who does not embarrass them.
- What they can do about it
- Walk away, and take the accreditation route and the campus with them, because the academy now sits on their estate rather than ours.
- How you find out you have lost them
- At the end of an agreement, or gradually, as placement numbers are quietly reduced.
-
The public hospitals we hire from
Neither- What they want
- To keep the consultants they trained and cannot replace.
- What they can do about it
- Nothing. They cannot outbid, they have no standing to object, and there is no forum in which the transfer is even recorded as an event.
- How you find out you have lost them
- Never, from here. The only way this becomes visible is if we measure it ourselves and publish it, which is the commitment made in the solution and the reason it was made.
Chapter 3 · And how long that takes to see
The ones who just stop coming
Take the referring clinician, because they are the clearest case and, in a market where the most trusted voice in the whole decision is already pointing at an airport, the most consequential.
A referrer who loses confidence does not resign. There is nothing to resign from. They simply send the next patient somewhere else, and the hospital receives exactly what it receives when that referrer happens to have had a quiet month, which is nothing.
The arithmetic of that is unforgiving and it runs the wrong way. The busier the source, the faster their silence is detectable. The quieter the source, the longer it takes. And a referral panel is mostly quiet sources, because that is the shape of every panel, so most of your relationships are in the rows where a departure takes half a year or a year to become a question anybody could ask.
Then the second number, which is why watching the total does not help. Losing five of the smallest referrers at once moves the headline by less than two per cent, which is well inside ordinary month to month variation. There is no chart on which this appears. There is no meeting at which it comes up. And by the time the trend is undeniable, the reason has been forgotten by the only people who knew it.
Which produces the one operational instruction on this page. Watch referral rate by source, not in aggregate, and treat a source going quiet as an event requiring a phone call rather than as data requiring more data.
Why the referrer matters so much in this particular market is on why we still fly.
| A referrer who sends | How long their silence looks normal | What you see in the meantime |
|---|---|---|
| 4 a month | 0.7 months before you could be 95% sure they had stopped rather than been quiet | A busy source going quiet is obvious within weeks, which is why the few large referrers are the only ones anybody manages. |
| 2 a month | 1.5 months before you could be 95% sure they had stopped rather than been quiet | Still visible inside a quarter, if somebody is looking at sources individually rather than at the total. |
| 1 a month | 3.0 months before you could be 95% sure they had stopped rather than been quiet | A quiet quarter is an ordinary quarter. Nobody investigates an ordinary quarter. |
| 0.5 a month | 6.0 months before you could be 95% sure they had stopped rather than been quiet | Two quarters of nothing from a source that often sends nothing for a quarter. There is no moment at which this becomes a question. |
| 0.25 a month | 12.0 months before you could be 95% sure they had stopped rather than been quiet | A year. By the time it is detectable the reason has been forgotten by the person who had it, and they have a new habit. |
66referrals a month across the whole panel, which is the only number anybody actually watches
1.9%the fall in that number if 5 of the smallest referrers leave at once, which is inside ordinary month to month variation
0of them will tell you, because none of them has resigned from anything
derived: Poisson arrivals at the stated rate per source. Months to notice is the point at which a run of zeros becomes unlikely enough to act on
Chapter 4 · Which is why you hear from the wrong people
Voice is expensive and exit is cheap
The asymmetry that makes all of this hard is that the two instruments cost their users very different amounts.
Speaking up costs something every time. A nurse raising a concern about a consultant's decision spends social capital and takes a risk. A family calling for a second opinion over the heads of the team treating their mother has to believe they are entitled to and that it will not rebound. A referrer telling a hospital that its discharge summaries are useless is picking a fight with people they will need again.
Leaving costs almost nothing, and it is final, and it requires no explanation to anybody.
So the default behaviour of a rational person with both instruments is to use the cheap one, which means an organisation hears from the people who cannot leave and loses the people who can, quietly, while its feedback channels report that things are broadly fine. Complaint volume is not a measure of how much is wrong. It is a measure of how many people still think saying something is worth the trouble.
The correction is not to collect more opinions. It is to make voice cheaper for the people who have exit, which is precisely what the escalation route on the family page is for and precisely why the volume of calls has to be read as health rather than as failure. A route that is free to use is the only thing that competes with a departure that is free to take.
The andon cord version of this argument, and the measurement rule that protects it, are on friends and family.
Chapter 5 · The host community
Obstruction is what is left
A group with no exit and no voice does not simply absorb. Some of them find a third instrument, and it is the only one available: get in the way.
The host community around a hospital site cannot take their custom elsewhere, because they were never customers. They have no formal channel, because none was built. What they have is proximity and numbers, which converts into a blocked access road, a stalled delivery, a site that cannot be worked, or a licence application that acquires local objections at exactly the wrong moment.
Inside the organisation that arrives as unreasonableness, and it is managed with security, lawyers and patience. That reading is comfortable and wrong. Obstruction is the predictable behaviour of people who were given nothing else, and responding to it as a security problem rather than as a channel problem guarantees it repeats, because nothing about the underlying position has changed.
The alternative is cheap and has to be done early. A named contact who answers. Employment commitments that are specific and checkable rather than warm. A published response time for the things that genuinely affect them, which are noise, traffic, water, waste and the mortuary. None of that is philanthropy and none of it is public relations. It is buying a channel so that the only available instrument stops being the one that closes the gate.
Chapter 6 · The obligation that has to be discharged differently
The two with neither, and why consultation cannot reach them
Two groups on the register have no exit, no voice and no route to obstruction. The patients in the catchment who cannot pay, and the public hospitals whose consultants this project intends to hire.
Both carry real costs from decisions taken here. Neither is in any relationship with the hospital, which is precisely why they have no instrument: exit requires being a customer, voice requires a channel, and obstruction requires proximity and a shared grievance. They have none of the three.
This matters because the standard organisational response to a stakeholder obligation is engagement, and engagement cannot reach them. There is nobody to consult. A community meeting will be attended by people who have a relationship with the hospital. The person who did not come because they could not pay is not at the meeting, and the consultant who has not yet been hired away is still at their post.
So the obligation has to be discharged by a different mechanism entirely, and there is only one that works: measure what happens to them and publish it, on a schedule set in advance, whether or not the number flatters us. That is what the clinician count commitment in the solution is, and this is the argument for why it is not a gesture. It is the substitute for a voice that does not exist, and it is the only one available.
It also explains why that commitment has to be made before the hospital opens rather than after. A baseline taken afterwards measures the world the hospital has already changed.
The commitment itself, and the debt it repays, are on the solution to instruction 01.
Chapter 7 · Three mechanisms, not one engagement plan
What to build, for each kind
Which reduces to three mechanisms rather than a single engagement plan, because the three groups cannot be reached by the same thing and an organisation that tries will succeed only with the group that was already talking to it.
An escalation route that is free to use, answered fast, and never appraised on how rarely it is pulled. The organisation that congratulates itself on low complaint numbers has built a system that punishes complaining and has no idea it has done so.
A baseline taken before opening, a number reported annually, and the bad years included. It is not engagement and it is not a substitute for consent. It is the only instrument that exists when the other party has none.
Chapter 8 · Three, published with it
Where this falls short
Three, and the first is the trap this page is most likely to lead somebody into.
Everything above argues for building voice for people who have exit. The easiest way to appear to do that is to run surveys, and surveys are answered by the people who have not left. A feedback system can report improving satisfaction for years while the underlying relationship drains, because the population being sampled is the population that stayed. The only reliable correction is to measure what people do rather than what they say, which is less flattering, considerably less comfortable, and the entire reason the referral rate by source belongs on a dashboard that somebody senior actually reads.
-
Building voice for people with exit is easy to fake
Falls short- Where it is weak
- Referrer surveys, exit interviews and patient feedback all produce data, and all of them systematically collect the opinions of people who have not left yet or who have no reason to be candid on the way out. The ones who went quietly are exactly the ones not answering.
- Who carries it if we are wrong
- The organisation, in a feedback system that reports satisfaction while the underlying relationship drains.
- What would settle it
- Measuring behaviour rather than opinion. Referral rate by source, not referrer satisfaction. Whether a consultant's leaving date clusters with a known event, not what they said at the interview.
-
The obligation to the people with neither cannot be consulted away
Irreducible- Where it is weak
- Two groups on this list have no exit and no voice, and both of them carry real costs from decisions taken here. No amount of engagement reaches them, because engagement requires a relationship and there is none.
- Who carries it if we are wrong
- Them, continuously, invisibly.
- What would settle it
- Nothing settles it. The only honest substitute for their voice is published measurement taken on their behalf and reported whether or not it flatters us, which is why the clinician count commitment is the load bearing promise in this whole study.
-
Obstruction is rational and treating it as unreasonable makes it worse
Falls short- Where it is weak
- A blocked gate reads inside the organisation as extortion or as a local difficulty to be managed. It is usually the predictable behaviour of people who were given no other instrument, and responding with security rather than with a channel guarantees a repeat.
- Who carries it if we are wrong
- The programme, in delay, and the relationship, permanently.
- What would settle it
- Giving the host community a real channel before the first event rather than after it, with something actually at stake in it. Employment commitments, a named contact and a published response time are cheap next to a week of blocked access.
The decision all of them have a stake in
Where to build, how to staff it, and who ends up carrying the cost of both.
Read instruction 01