Strategic mandate and decision boundaries · Lesson 7

Allocate portfolio outcomes across unequal service units

17 min read

Lesson objectives

  • Compare allocations that meet the aggregate objective but differ in unit outcomes and service floors.
  • Test allocation feasibility, timing and robustness before committing resources.
Reading tools

An organisation wants to reduce service delays across five regions. One proposal assigns the same percentage reduction to every region; another divides funding by population. Both can sound fair before differences in need are examined. A region with stable service may need a small improvement, another starts with severe delays, and a third faces a short seasonal surge unlike the rest of its year. The portfolio decision is to define a shared outcome and distribute commitments in a way that explains these differences. The leader separates an entitlement to a minimum service level from competition for additional improvement. They then test what each unit can deliver with its resources and whether the allocation leaves a group of users behind an improving aggregate. All figures in this lesson are fictional training examples, not an allocation rule prescribed for a public or private organisation.

Choose the unit of need before choosing the unit of funding. Need might mean requests delayed beyond an acceptable limit, people unable to access the service, or high-risk cases requiring faster action. These are different populations; population size alone does not always estimate them. Also distinguish observed need from registered demand. Weak registration in a remote region may depress recorded demand even though access is the main problem. State what is counted, the observation period, the source and who is absent from the record. If the missing population remains unknown, commission a short verification task rather than treating it as zero. Do not substitute an invented estimate presented as an observed fact.

Present three distinct layers of the decision. The first is the service floor promised to each eligible unit or group. The second is the additional improvement the portfolio can purchase with available resources. The third is the distributional effect that may warrant an explicit constraint or weighting. The UK Green Book explains the importance of examining option impacts, including how benefits and costs are distributed among affected groups. That perspective structures the allocation question here; it does not transfer a UK legal obligation to another organisation. Local limits require approval from the appropriate authority, with their source and date recorded. Where no limit exists, identify the suggested rule as a proposed value judgement awaiting a decision, rather than an existing obligation discovered through arithmetic.

A floor differs from a weight. A floor rules out an allocation that leaves a unit below a specified service level, even when that allocation produces a larger total outcome. A weight increases the importance of some units' outcomes in comparison, but may still permit poor performance in one unit when weighted gains elsewhere compensate. Ask which the decision-maker intends. Saying that disadvantaged regions deserve priority does not specify whether a minimum must first be guaranteed or whether their gains should simply carry more weight. Present both interpretations with their costs and outcomes, and record the approved choice. Do not select an undisclosed weight merely to make the calculation yield your preferred allocation.

This calculation compares marginal improvement: the contribution of the next team after teams already allocated. Multiplying the first team's effect by ten is unjustified unless ten independent opportunities of that value exist. Interventions may overlap, equipment constraints may emerge, or readily improvable work may run out. Ask the unit owner for the effect of each increment, its evidence and its conditions. If a team cannot deliver without preparatory training, show the complete package, its cost and the waiting time. Where team durations differ, use a common effort unit or disclose that packages differ. Matching team labels does not make their effective capacity equal.

Test whether the service floor is feasible before optimising allocations within it. Inland has two hundred and fifty delayed cases out of five hundred. Reaching 35% requires at least seventy-five fewer delays, requiring three teams under the estimate. That resource commitment precedes allocation of the remainder. A requirement of 20% in every unit instead requires one hundred and fifty fewer inland delays. Three teams cannot achieve this under the stated marginal effects. Calculate the contribution of further teams or find a different intervention before declaring the plan feasible. If the floor cannot be met within the resource ceiling, escalate the conflict with clear alternatives: more capacity, a redesigned intervention or an explicitly changed objective. Do not submit an arithmetically infeasible plan and expect pressure on staff to close the gap.

Examine seasonality in the period when improvement will actually arrive. One hundred delayed cases during a two-week season may require immediate capacity; the same number spread across a year may permit gradual intervention. If teams arrive after the season, a theoretical annual reduction does not relieve the current bottleneck. Show demand timing, intervention availability and the lag until outcomes appear. Separate service provision in the current season from improvement for the next one. Borrowing a team temporarily might address the first, while a booking-system change addresses the second. Counting both as if they help the same people at the same time exaggerates portfolio impact.

Funding conditions, staff qualifications or asset locations may constrain allocations. Do not represent eligibility restrictions as preference weights: first check whether the resource can perform the task. A team available on the coast may not immediately work in the mountains if different authorisation or unbudgeted accommodation is necessary. Identify transferable resources and transfer conditions, then construct feasible allocations. Distinguish the transfer cost borne by the portfolio from the loss borne by the lending unit when its team leaves. Where a resource comes from another service, include the consequences for that service's users. Improvement in two units does not justify hiding deterioration in a third outside the dashboard's scope.

Comparison also needs a clear spending baseline. The worst-performing unit may receive less funding per request, or its higher cost may reflect dispersed locations rather than weaker efficiency. Separate unavoidable environmental requirements from operating errors where the evidence permits. This distinction does not remove accountability; it identifies what can change. Transferring a resource, redesigning a process and introducing a mobile service solve different problems. If the data cannot separate causes, present a range or conditional alternatives. A short diagnostic investigation that rules out a decisive explanation may be preferable to a large expansion based on a league table that does not explain differences between sites.

Define how the allocation will be reviewed after implementation begins. If inland team impact proves weaker than expected, immediately moving teams to the coast might improve the aggregate while breaching the local floor. Show the consequences of retaining teams, transferring them and correcting the cause of weak performance. Lower impact could be a temporary learning effect or evidence that the design does not fit the setting. Set the review date around when an effect can reasonably emerge, rather than solely around a standing meeting. The UK government project-delivery standard describes benefits monitoring and ownership. This lesson uses that principle to assign responsibility for each result; it does not claim that the standard prescribes this allocation formula.

Do not treat units as interchangeable containers when assigning responsibility. Each local outcome owner verifies demand and delay data; the portfolio manager shows combined impact and cross-unit conflicts; the committee decides changes exceeding delegation. Define a simple exception path: who reports a threatened local floor, what evidence is needed and when a decision must occur. The committee needs the affected users and timing, rather than only a status colour. Retain both the original allocation and amendments so later reviewers can explain why a unit's commitment changed. Otherwise centrally imposed resource transfers may later be misrepresented as failures of local execution.

Include users' views in understanding distribution while keeping decision authority clear. Users in a remote area might prefer a reliable weekly appointment to a daily office that frequently cancels. Do not reduce the distinction to one delay rate if it hides the kind of access they value. Use focused consultation to understand need and connect findings to a feasible design. Participation does not mean the last speaker represents everyone. Document who participated, who could not be reached, how the information affected the design and what remains unknown. Listening to regions then becomes evidence informing a choice, rather than a slogan substituting for need measurement or overriding a specific financial delegation.

Start the final decision note with an outcome users can recognise, then show its distribution. For each unit, show the baseline, period demand, approved floor, expected improvement, resources and leading risk. Add a total recalculated using demand volumes as weights, not an unweighted mean of site percentages when volumes differ. Include a feasible alternative that exposes the price of the choice. It might produce the same aggregate but leave one unit worse off, or strengthen a local floor at greater cost. This explains the decision without pretending that one allocation is entirely value-free. Allocation includes an explicit value judgement; arithmetic clarifies its consequences rather than replacing the person authorised to make it.

Apply the lesson to one current portfolio decision. Find an aggregate objective that conceals different starting points or timing of need, then restate it using users and actual resources. If the current allocation satisfies limits and has a clear explanation, retain it and record why it remains suitable; the lesson does not demand change simply to demonstrate activity. If a conflict appears, show its size, a feasible alternative and the authority that must decide. The professional output is an allocation whose consequences can be calculated, whose owner can be held accountable and which can be revised as conditions change, with explicit recognition of remaining uncertainty about need and intervention impact.

Test allocation robustness when resource impact is a range rather than a single number. The central estimate is the best available reading, but it does not guarantee the floor under every condition. Establish whether the committee seeks an acceptable probability of achievement or protection against weaker impact within a stated range. Do not infer a probability from optimistic and conservative endpoints alone: probability requires a distribution or further evidence. An allocation can be presented as meeting the floor centrally but failing under the conservative case, provided a response is explicit. This prevents risk from hiding behind decimal precision and prevents arbitrary contingency unrelated to the size of the shortfall.

Finally, examine who bears improvement costs and who receives the outcome. Central management may fund an intervention while the local unit absorbs training hours and transition backlog. If those burdens are hidden, a low-capacity unit may appear to reject a free improvement. Include participation time in its resource constraint and specify which work will be postponed. Do not count a benefit twice because both central and local owners report it: the total prevented delays remain one outcome. Linking these records reconciles resources and results, turning local objections into testable information rather than a generic accusation of resistance to change.

Lesson summary

The lesson distinguishes user need from registered demand and a local floor from a preference weight. It calculates demand-weighted rates and tests capacity, seasonality, transfer conditions and uncertainty. The result is an explicit allocation with stated authority and distributional effects, revisable when evidence changes.