Claire MercierChief Financial Officer · Essays & perspective
← All essays

Finance transformation3 min read

Capital allocation needs a common language

Two investment cases can appear comparable while using different assumptions about risk, time and resource demands.

A common appraisal framework makes these differences visible without pretending that every strategic benefit can be reduced to one number.

I ask sponsors to explain downside exposure, cash requirements and the milestones that would justify further commitment. Accountability remains with the business after approval. A post-investment review should test the original logic, not simply confirm that the budget was spent.

Different appraisal languages persist because each part of a business develops its own conventions over time. Capital projects in operations, technology investments and marketing programmes are often assessed by different teams, with different horizons and different views of risk. Proposals labelled strategic sometimes claim an exemption from scrutiny altogether. Sponsors learn the system quickly and present their cases in whichever terms the local process rewards. None of this is dishonest, but it leaves the investment committee comparing documents that look alike while resting on quite different foundations.

My preferred framework is short and consistent. Every proposal states its cash profile over time, the assumptions that drive its returns, a credible downside case, how reversible the commitment is and the milestones at which further funding would be released. It also states what else the money could do, so the committee is reminded that approval is always a choice between alternatives. Alongside the sponsor's case, finance prepares a brief independent note that tests the key assumptions. The sponsor sees that note in advance, which keeps the challenge constructive rather than theatrical.

Consider an investment committee comparing a plant expansion with a new digital platform for customers. The expansion has clear cash returns and a long asset life. The platform claims strategic benefits that are harder to quantify, such as retention and future pricing flexibility. A common language does not force both into the same return metric. It asks each to state what must be true, when the evidence will arrive and how much is committed before then. The platform might then be approved in stages, with each release of funding tied to a test of its central claim.

A familiar criticism is that common frameworks become bureaucratic and discourage good ideas, especially small or unusual ones. I take that seriously. The depth of analysis should be proportionate to the size and irreversibility of the commitment, so a modest experiment needs far less than a major acquisition. Another concern is that sponsors will simply learn to game a standard template. The post-investment review is the best defence here. When original claims are revisited openly and consistently, the incentive to inflate them weakens, and the quality of proposals tends to rise without any further instruction.

The framework only holds if the people applying it accept it for themselves. Committee members must bring their own proposals in the same form and accept the same challenge they direct at others. Finance must acknowledge when its own assumptions, about inflation, financing costs or timing, proved wrong. The board can help by asking periodically how the overall portfolio is balanced between protecting the existing business and building new options, a question that individual approvals rarely answer. When those habits are in place, the language becomes part of how the organisation thinks rather than a form it completes.

A common language also protects the organisation from its most persuasive advocates. When every proposal states its returns, its risks and its reversibility in the same terms, the investment committee can compare substance rather than presentation. The capital that follows is better placed, and the business leaders who earn it know exactly what they have committed to deliver.