The Tranche Boundary That Shows No Benefit
A utility is two years into a five-year programme to install 1.2 million smart meters and a new billing platform. Tranche 1 (years 1–2) delivers the billing platform and 200,000 meters. The £38m annual benefit only begins once more than 60% of customers are on the new platform, planned for tranche 3. At the tranche boundary review the finance director notes cumulative spend of £61m against zero benefit realised, observes that the billing platform alone would deliver about £9m a year, and proposes cancelling the meter rollout and running the platform as a standalone project. Field engineers report installation costs running 18% over estimate because of access problems in older housing stock.
1. How should the programme manager respond to the finance director's proposal?
Read the worked answer
Separate the legitimate observation from the invalid conclusion.
The observation is correct and serious: a tranche that realises nothing is a design flaw in the tranche plan, not bad luck. The conclusion is wrong, because it assumes benefits are additive when they are a step change. The £38m depends on scale; split out, the platform yields £9m - which, against £61m already spent plus remaining cost, probably fails its own business case. The finance director has identified a problem and proposed a cure worse than the disease.
So the response is to accept the critique and fix the tranche design:
- Re-profile tranches so a measurable benefit lands inside tranche 1 or 2 - for example, targeting the 60% threshold in two high-density urban regions first, or defining intermediate benefits (reduced estimated billing, fewer billing queries, lower bad debt) that the platform alone genuinely produces.
- Re-test the 60% assumption. It may be a planning convenience rather than a hard technical limit; if benefits begin at 35%, tranche 2 becomes a benefit tranche and the whole argument dissolves.
- If no intermediate benefit can be identified, say so plainly: the programme's case rests entirely on the end state, and the board must fund it on that basis with its eyes open.
What the programme manager must not do is defend the current plan on the grounds that it was approved two years ago. A tranche boundary exists precisely to re-test the business case.
2. What should the programme manager do about the 18% installation overrun?
Read the worked answer
Diagnose before responding - the overrun has three possible homes and they need different treatments:
- Estimate error (access issues were always there and were under-surveyed): re-baseline the tranche plan and the business case cost line, and fix the survey method before scaling. Cheapest fix, but it means telling the board the number has moved.
- Delivery inefficiency (crews idle, scheduling poor): a delivery problem, addressed through the tranche plan and supplier performance, not through the business case.
- Blueprint constraint (the specified meter or installation method is unsuitable for older housing stock): the most serious, because it means the capability as designed cannot be delivered across the whole estate. That is a blueprint change, requiring board approval, and may need a second meter variant or an alternative installation approach.
With 1.2 million installations ahead, the programme manager should treat the 18% as a signal from tranche 1 to tranche 2, not as a variance to be absorbed. Concretely: segment the housing stock, calculate the overrun by segment, and check whether the older-stock segment alone is running far higher than 18% - if so, the average is hiding a structural problem. Then re-forecast total rollout cost and re-test the benefit case at the new cost before agreeing the tranche 2 plan, so the finance director gets a revised number rather than an assurance.