I used to think that the main intellectual work in an investment paper was challenging each proposal. If every sponsor could defend the assumptions, and finance could reconcile the arithmetic, the combined account seemed reasonably secure. That view treated a portfolio as a collection of independent objects. In practice, programmes often depend on the same operating change, compete for the same resource or describe different stages of a single improvement. The relationships between them can matter as much as their separate cases.
My clearest lesson came from a 2018 proposal at Qevzenti. Three programmes sought ₹41 crore in combined investment, with annual prospective benefits shown as ₹36 crore. I had supported the paper as head of strategy. A subsequent review identified ₹9 crore attributed more than once. Removing that overlap produced a ₹27 crore nonduplicative estimate for the full three-programme model. Neither figure was realised profit. The failure was in the account I had helped assemble, not evidence that an accounting ledger had necessarily recorded an error.
An enabling contribution is not another enterprise receipt
A sponsor's first response was understandable: without the programme, the improvement would not happen, so its case should include the improvement. Another sponsor could make a similar argument about a different dependency. Both might be right about necessity. It does not follow that the enterprise receives the benefit once for each necessary contribution. A decision paper must be able to recognise enabling work without manufacturing another economic receipt for every team involved.
I now begin with the operating change being described. Which activity becomes different? What expense ceases, what capacity becomes available, or what result could improve? Who would be able to observe that change? These questions make the proposed benefit more concrete than a label such as efficiency. They also help colleagues identify when two programmes are pointing to the same underlying event, even if they use different terminology or place it in different reporting periods.
The next question concerns the counterfactual. A programme may claim an avoided future expense, while another claims the same amount as a reduction from the current cost base. Those are different descriptions and may rely on incompatible starting points. The strategy function should not settle the matter through elegant language. Finance and operating owners need to agree what is being compared and which costs remain under each alternative. The purpose is an intelligible decision, not an allocation contest between sponsors.
This can be uncomfortable because the benefit attached to a programme often carries organisational status. Removing an amount may feel like questioning the team's usefulness. I try to separate the usefulness of the work from the accounting of its contribution. A programme can be indispensable to an improvement without being entitled to repeat the whole benefit in an enterprise total. Showing the dependency explicitly is usually a better representation of its importance than maintaining a larger but misleading number.
The selected combination needs its own case
Our corrected ₹27 crore estimate concerned all three proposed programmes together. The eventual authorised selection was two programmes with ₹28 crore proposed capital expenditure, while the third was deferred. It would have been a second mistake to place the full proposal's corrected benefits beside that smaller investment amount and imply a better return. A change in selection changes the system being evaluated. Some benefits may survive, some may diminish and some may depend on the omitted work.
That is why I want the decision record to preserve the identity of the combination. A board should be able to tell whether it is looking at the original portfolio, a reduced version or a staged approach whose later elements remain unapproved. A single benefits column that survives every revision can make distinct choices look interchangeable. The more persuasive the resulting ratio appears, the more important it is to check whether its numerator and denominator describe the same option.
There is a practical tension here. Teams cannot model every conceivable combination at equal depth. Excessive analysis can delay an otherwise sensible decision and consume the very expertise needed for delivery. I favour identifying the material dependencies first, then developing the few combinations that change the choice. The standard is not perfect foresight. It is sufficient clarity for the approving body to understand what it is committing to and which assumptions would need another decision.
Timing belongs in that account as well. Benefits described as annual may begin in different years or require a period of transition during which costs continue. A sum of annual run-rate estimates is not automatically a first-year outcome. I ask for the relevant period to remain visible even when a presentation is simplified. Finance owns the financial interpretation; strategy should help ensure that simplification does not quietly change the proposition being put to the board.
Correction belongs beside the original recommendation
We spent ₹38 lakh on the review of definitions, dependencies and alternatives. That expenditure was a consequence of doing the work again properly. I would not portray it as an investment that had already paid for itself through the numerical correction. The difference between ₹36 crore and ₹27 crore was a change in a prospective estimate, not money found in an account. Recognising that distinction prevents the repair from becoming another inflated success story.
I also believe the strategy leader should remain visible in the account of the mistake. It would have been easy to say that sponsors supplied inconsistent figures. They did supply the cases, but I was responsible for the combined recommendation and had defined the questions the planning team was asked to examine. The omission sat in that design. A useful learning record identifies the decision-maker's own assumption, the challenge that exposed it and the work required to produce a better recommendation.
This does not mean every proposal should begin with a warning about double counting. It means the review should follow the enterprise change rather than the organisation chart. The board needs to know what can happen together, what depends on something else and what becomes unavailable when a different combination is selected. In this fictional case, the better strategy paper was smaller in its claims and more precise about the choice. That is a standard I would carry into another investment discussion: make the relationships visible before allowing the total to speak for itself.