Illustrative · Kestrel Pathways is a constructed operator, not a real provider. Its shape, scale and contract mix are drawn from how the UK employability sector actually works, and every figure is indicative unless it carries a source.
Case study 02
Kestrel Pathways
An employment, health and skills provider of 2,100 people, living on payment by results. It does serious work and it is paid for a proxy of that work. This study examines what that gap does to the organisation, long before anyone proposes a solution.
Kestrel Pathways helps people into work. That sentence covers a national employment programme, a health-led service commissioned jointly with an integrated care board, a set of skills courses, and a smaller contract working with people leaving custody. Seventy four thousand referrals a year, two thousand one hundred staff, a hundred and eighty places where somebody sits down with somebody else and tries to change what happens next.
It is paid, mostly, when a person starts a job and stays in it. That is not a criticism of the commissioner: an outcome that can be audited is the only kind a public body can safely buy, and the alternative, paying for effort, funded a generation of programmes that produced activity and nothing else. But the funded measure is a proxy for the purpose, and the distance between a proxy and a purpose is where organisations quietly deform.
This study sets out the company, the people in it, the person it exists for and the way it gets paid, before any problem is stated and long before anything is solved. The first instruction follows once the ground is laid.
Chapter 1 · Attributes
What the company is
A provider of this size is three businesses wearing one name. There is the delivery business, which is advisers and rooms and caseloads. There is the bid business, which is a permanent function because contracts end and the next one has to be won two years before it starts. And there is the performance business, which exists because the money arrives against evidence, so somebody has to collect, validate and defend the evidence for every claim.
The last of those is invisible from outside and enormous from inside. A job outcome is not a job; it is a job plus proof of a job, of a kind an auditor will accept eighteen months later. That distinction generates whole departments, and it is the reason a company like this employs seven hundred people who never meet a participant.
- Legal form
- Private companyOwned by a small group of institutional investors, with a minority holding retained by the founders.
- Revenue
- £135m a yearIndicative. Ninety four percent from public contracts, the rest from employer-funded and local authority work.
- Staff
- 2,100About 1,400 of them are advisers, health specialists or employer-facing. The rest are performance, compliance, data, bid and central functions.
- Locations
- 180 delivery sitesTwenty six leased offices and the rest outreach: libraries, job centres, GP surgeries, community centres, one room a week at a time.
- People supported
- 74,000 a yearReferrals, not individuals. A meaningful minority are the same people returning after an outcome was already claimed for them.
- Contract mix
- 4 major, 11 minorEmployment, health-led employment support, skills, and justice-linked services commissioned by three different departments with three different theories of the same person.
- Payment mix
- 55% outcome, 30% service fee, 15% grantIndicative. The outcome share rose at each re-procurement, which is the single most consequential fact about this company.
- Revenue concentration
- 61% in three contractsAll three re-procure within twenty months of each other. Nothing else in this study matters as much as that sentence.
Chapter 2 · The actor model
Who is in the room
Two actors and a payer. The first actor is the participant, who carries every cost of the system directly. The second is not a party but a scale: the provider runs from a single adviser at one end to the group board at the other, by way of the team, the office and the region. What a critic calls "the company" and what a participant calls "my adviser" are the same actor at two ends of one line.
The law that falls out of it is the same as in any organisation of this kind, and it is worth stating plainly. The ability to see what something costs the participant falls as you move up the scale, while control over the decision that caused it rises. The two cross somewhere in the middle. Below the crossing people see and cannot act. Above it they act and cannot see. Nobody in this picture is behaving badly.
The payer sits outside the scale and inverts it completely. The commissioner holds the most control of anyone here, because it writes the rules, and sees the least of anyone here, because it meets the participant through a monthly data return. Every behaviour on the chart below it is a rational response to something written by a party that will never watch it happen.
-
The participant
Actor one- Its version of the goal
- Set out in full on the participant page. In short: money that is not frightening, work I can keep, no harm to my benefit while I try, and to be treated as an adult who has done this before.
- Distance from the patient
- Zero. Carries every cost of the system directly and holds almost no lever over any of it.
-
The adviser
Actor two, smallest- Its version of the goal
- Get this person into work that lasts, inside a caseload that makes that arithmetically difficult.
- Distance from the patient
- One conversation. Sees the whole of the person and controls almost nothing about the offer.
-
The office manager
Actor two- Its version of the goal
- Hit the site target without losing the team, and explain the gap when it is missed.
- Distance from the patient
- One floor. Sees the caseload in aggregate and the individuals only when they escalate.
-
The regional lead
Actor two- Its version of the goal
- Hold contract performance across thirty sites, and keep the commissioner relationship intact.
- Distance from the patient
- A county. Sees a dashboard and a monthly exception report.
-
The group
Actor two, largest- Its version of the goal
- Win the next re-procurement, protect the margin, and keep the balance sheet able to carry an outcome-funded bid.
- Distance from the patient
- A head office. Sees money in full and the conversation almost not at all.
-
The commissioner
The payer, outside the scale- Its version of the goal
- Buy outcomes that can be audited, at a price that survives scrutiny, from providers who will still exist in four years.
- Distance from the patient
- A department. Writes the rules that produce every behaviour on this chart, and meets the participant least of anyone on it.
Chapter 3 · Current position
The position it is in
Three contracts carry sixty one percent of the revenue and all three re-procure inside twenty months of each other. That is the fact that shapes every decision in the building, including ones that appear to be about participants. A company facing that cliff behaves differently from one that is not, and pretending otherwise would make this study useless.
Margins are thin and the outcome share has risen at each re-procurement, which transfers risk from the commissioner to the provider. The provider carries it in two ways: by holding working capital against outcomes that may not arrive, and by making choices about whom to work with hardest. The second of those is the subject of the first instruction.
Meanwhile the work itself has got harder. The people referred now are further from work than the people referred five years ago, because the ones closest to work were helped by the last programme, or by the labour market, or by themselves. The price per outcome has not moved in proportion. Everybody in the sector knows this and the bidding continues anyway, which is itself a fact about how the market clears.
Chapter 4 · Boundary
What this study will not do
It will not argue that payment by results is wrong. It is a reasonable answer to a real problem, and the programmes that preceded it were worse in ways that are easy to forget. It will not blame advisers for behaviour the funding model rewards, and it will not blame commissioners for wanting evidence.
What it will do is state the mechanisms plainly, name who carries each cost, and say where a number does not exist rather than inventing one. Then it will state one instruction at a time, publish the criteria and the comparator set before any answer is scored, and let the answer be graded against them in public.