How do we stop parking the third furthest from work?

Kestrel Pathways has a value that says it does not park people, and a funding model that pays it to. The question is what the company should do about that, given that it cannot change the funding model, cannot afford to lose the contract, and has eighteen months until the three largest contracts re-procure.

Stated precisely: what mechanism, inside the provider gift, would measurably increase the effort and the outcomes for the third of the caseload assessed as furthest from work, without reducing overall contract performance below the level at which the contract is at risk, and would survive a bad quarter without being suspended?

It is a brief for a decision, not for a campaign. An answer that argues the commissioner should change the rules has answered a question that was not asked, however right it is.

The first column is what the contract asks for and the company reports. The second is what would be needed to know whether the company is parking anybody. The third is what the difference costs.

Note that the company cannot currently prove it is parking people, and equally cannot prove it is not. That is the honest starting position, and any answer that opens by asserting the scale of the problem is inventing a number.

Five alternatives will be scored, and they were chosen before any weighting was set. They are: do nothing and keep the value as an aspiration; reallocate the most experienced advisers to the bottom tertile; cap caseloads for advisers holding the hardest cases; create a separate specialist team for the furthest from work; and publish outcome and effort data by tertile, internally first and then externally, with nothing else changed.

The comparator set contains at least one option somebody credible would actually argue for, which is the fourth, and it contains doing nothing, because a brief that excludes the status quo has already decided. Two options were considered and excluded, and they are named as exclusions rather than omitted: refusing referrals of the hardest cases, which is excluded as contrary to the contract and to the point of the company, and lobbying the commissioner for differential pricing, which is excluded from this instruction because it is not inside the provider gift, though it remains the single most effective structural answer and the study says so elsewhere.

Five criteria, weighted, with the method of scoring named for each. The threshold: an option scoring below sixty out of a hundred is not recommended regardless of how it compares with the others, because the alternatives on this question include doing nothing honestly, which at least does not add cost.

The fifth criterion carries five percent and functions as a veto. An option whose effect cannot be measured by anybody other than the people it judges is not scored at all.

Contract performance may not fall below the level at which the commissioner may exercise its remedies. That floor is a number in the contract, it is known, and no answer may cross it on the argument that the cause is virtuous.

Adviser headcount is fixed for the current financial year. Answers may reallocate people and may change what they do; they may not assume new hires.

Nothing may be done that requires the participant to be told less than the truth about what is being recorded, reported, or decided about them. That is a constraint rather than a criterion, so it is not tradeable against performance.

And the change must be describable to the commissioner without embarrassment. Any mechanism that only works if the payer does not know about it is not an answer, it is an exposure.

Four mechanisms, each named with how it works, who carries it, and what it would take to measure. The third one has no measure and says so, because assigning it a probability would dress a judgement about a relationship as an analysis.

A recommendation, scored against the five criteria, with the score shown rather than asserted, and the comparison against all five comparators including doing nothing.

A baseline: what the bottom tertile receives and achieves today, measured before anything changes, because a change without a baseline cannot be evaluated later and will be claimed as a success by whoever is still there.

A structural test: an explanation of what holds the change in place when the quarter goes badly and attention moves, and what specifically would be observed if it were quietly abandoned.

A cost, as a range, with its basis. And a named owner, at a level that has both the sight and the control, which the actor model says is roughly the regional lead and certainly not the board.

The answer will be published at the solution page for this instruction and graded against exactly these criteria, with the weights unchanged from this page.