Written to Win, Not to Manage: Why the Benefits Case Dies at Approval
Ambition and fiction are not the same failing, and they do not have the same cure.
The Document That Did Its Job Too Well
Somewhere inside every approved programme sits a benefits profile that has done exactly one job and will never be asked to do another. It was written to release the money. It released the money. And on the morning after approval it quietly becomes what it was always destined to become — a filed artefact, opened perhaps once a year at a portfolio review by someone reading its figures aloud for the first time, unable to say whether a single one of them ever arrived.
I have watched this pattern repeat across enough programmes to stop treating it as an accident of execution. The benefits case is not forgotten because the people running the programme are careless or unserious. It is forgotten because it was authored for a purpose that expired the moment the investment board said yes. We ask one document to be two irreconcilable things at once — a bid and a plan — and it performs superbly as the first while quietly failing as the second. Then we express surprise, year upon year, portfolio review upon portfolio review, that the value we were promised never seems to turn up.
The benefits case fails at realisation precisely because it succeeds at approval. The very qualities that make it a persuasive bid — optimism, aggregation, and a careful distance from any named owner — are the qualities that make it worthless as an instrument of management.
What the Bid Rewards
Consider a programme approved on the strength of eight million pounds a year in benefits. Five of those millions are described as productivity improvements: roughly half a day a week handed back to some four hundred staff, converted into money at an average loaded salary and totalled into a single confident line on the summary page. The number is not a lie, exactly. Every assumption beneath it is defensible taken on its own. But at no point between the business case and the benefits profile did anyone ask the one question that decides whether the benefit is real: when those half-days are returned, whose budget actually falls?
The answer, almost always, is nobody’s. The recovered time is genuine and the saving is fictional, because a saving that is never removed from a cost line is not a saving at all — it is a hope with a spreadsheet attached. Four hundred people each a little less busy is not five million pounds; it is four hundred people each a little less busy. To become money it would have to be consolidated, and consolidation means fewer posts, or reallocated work, or a service that visibly does more with the same headcount. None of that is planned, because none of it is needed to win approval. The line only has to survive the investment board, and it does.
This is not a failure of individuals but of the incentives we have built around them. Every portfolio I have worked within operates under capital rationing — a fixed pot and more claims on it than it can hold. In that setting, programmes do not merely make their case; they compete. And the bid that wins is reliably the one that promises the most value per pound committed. That selection pressure is relentless, and it selects for optimism with the efficiency of a market. A team that presents cautious, fully-owned, conservatively-attributed numbers loses the money to the team next door that does not. Over enough cycles, the organisation trains its most capable people to write benefits cases that are optimised for approval and structurally unfit for management — and then blames them for the gap.
The Aggregation That Hides the Owner
There is a second, quieter mechanism at work, and it hides inside the arithmetic. To make a compelling bid, benefits are summed. Dozens of small, specific, ownable effects — a shorter handling time here, a reduced error rate there, a licence no longer needed — are rolled upward into a handful of confident totals, because a board cannot weigh forty granular claims but can weigh one number. The aggregation is rational at the moment of persuasion and catastrophic for everything after it. A total has no owner. You cannot hold anyone to eight million pounds; you can only hold someone to the specific, decomposed line they agreed to deliver. The instant the granular effects are summed away, the accountability is summed away with them, and what remains is a figure that belongs to the programme in the abstract and to no manager in particular. When realisation time comes, there is no one to ask, because the number was never anyone’s to begin with. This is why re-decomposition — breaking the headline back down into lines a named person will sign for — is not administrative tidying. It is the act that decides whether a benefit is governable at all.
The Objection Worth Taking Seriously
There is a serious counter-argument, and it deserves better than the caricature it usually receives. It runs like this: a business case must lean optimistic, because investment is a contest for scarce capital and unrelieved caution simply hands the future to whoever is willing to promise more. A degree of optimism bias, on this view, is not a defect but the necessary grease of the whole system. Punish every over-claim and you do not get honest cases; you get paralysis, and the good programmes starve alongside the bad.
I have some sympathy with this, and none at all with the pretence that benefits cases could be made coldly accurate if only people were more disciplined. Forecasting value years out is genuinely hard, and a certain reach is legitimate — even necessary. But the argument proves less than it claims. The problem in front of us is not that the numbers are ambitious. It is that no one is left standing behind them once the cheque has cleared. Ambition and fiction are not the same failing, and they do not have the same cure. The remedy for optimism is not compulsory pessimism; it is accountability that outlives the approval it was written to secure.
“A benefit with no owner after approval is not a benefit. It is a decoration on a document that has already done its only job.”
What Portfolio Leadership Can Actually Do
None of this requires a new methodology, and I distrust the reflex that reaches for one. It requires the portfolio’s leadership to stop accepting the bid as if it were a plan. Four moves, in ascending order of discomfort, do most of the work.
- Refuse to let one document serve two masters. The case that wins the funding and the profile that governs the delivery should not be the same artefact. At approval, force a deliberate translation: the persuasive, aggregated bid is rewritten into a management instrument in which every number is decomposed to the level at which someone could actually be held to it. If a benefit cannot survive that translation, it was never a benefit — it was rhetoric.
- Attach every benefit to a named budget line and a named owner. Not the programme’s budget — the receiving business’s budget. A productivity saving must point at the cost line it will reduce and the manager who will have to run with less. Cash-releasing benefits that cannot name the line they release are reclassified on the spot as non-cash, and counted differently. This single discipline collapses most inflated cases without a single argument about assumptions.
- Re-baseline at the point of approval, not before delivery. The world the business case described is months old by the time the money is granted. Approval should trigger a fresh baseline against which realisation will actually be judged, owned by the receiving business rather than the programme — because the programme has every incentive to remember the flattering starting point and none to correct it.
- Make the portfolio review a place where benefits are removed. A review that can only add and track benefits, never strike them out, is an accounting fiction. The most valuable act a portfolio board can perform is to formally de-claim a benefit that has not materialised and will not — to say, in the minutes, that five million pounds was never going to appear and the plans that depend on it must change. An organisation that cannot bring itself to do this is not managing benefits. It is curating a museum of things it once hoped were true.
The uncomfortable truth beneath all four is that the benefits case, as most organisations use it, is an instrument of persuasion that we have quietly agreed to treat as an instrument of truth. There is nothing wrong with persuasion; capital has to be allocated and someone has to make the argument. The error is forgetting, the moment the money lands, that persuasion is what the document was for. The programmes that actually realise their benefits are not the ones with the most rigorous forecasts. They are the ones whose leadership keeps asking, long after everyone else has moved on, the awkward question the bid was carefully constructed never to raise: whose budget falls, and when?