The five part structure
Problem and cost of doing nothing: what hurts today, in a number, and what it compounds to if ignored, because the do-nothing baseline is the real competitor to every proposal. Options considered: two or three, including the cheap one, so the decider sees you chose rather than pitched. Recommendation with costs: money, named people’s time, and duration to the first gate. Expected benefit: the committed number this moves, by how much, by when, and how it will be measured. Risks and kill criteria: how this plausibly fails and what evidence would mean stopping.
Two pages forces the discipline. If the case cannot survive compression, the thinking is not done.
Honest numbers at small scale
SMB business cases fail through false precision more than through optimism: a spreadsheet projecting benefits to the dollar over three years communicates confidence the data cannot support, and experienced deciders discount it to zero. State ranges and the assumption that drives them: this saves roughly one day a week of one named role, worth this range annually, and the assumption to test in month one is such and such. A range with a named driver invites the useful argument, about the assumption, instead of the useless one, about the arithmetic.
Payback period usually serves a small company better than multi-year ROI: cash cycles are short, and a case that pays back inside a year survives most forecasting errors.
Benefits realization: the loop almost nobody closes
The project shipped, the team moved on, and nobody ever checked whether the day a week got saved. Closing the loop is one calendar entry: at the interval the case named, the sponsor spends fifteen minutes comparing promised benefit to observed reality, and the result is logged with the original case. The direct value is occasional course correction; the compounding value is cultural, because estimates that will be checked are made differently, and the company’s casework gets honest within a few cycles.
Cases, gates and the benefits check, on one record
The PMO Desk carries business cases as structured records tied to the commitment gate: the five parts, the kill criteria, and the benefit with its check date. The check surfaces on My Day when it comes due, months later, when memory alone would have let it slide, and the result logs beside the original promise.
Velora drafts the case from a plain description and stress-tests the benefit logic before the sponsor sees it. Decision quality as a habit, not a heroic effort, from the Plus membership on the workspace for small and midsize operators.
Frequently asked questions
How long should a business case be?
Two pages: problem and do-nothing cost, options, recommendation with costs, expected benefit with a measurement plan, risks with kill criteria. Detail beyond that is appendix material.
What financial measure fits small company cases?
Payback period, usually. It matches short cash cycles and stays meaningful under estimation error. Multi-year ROI models imply precision SMB data rarely supports.
What is benefits realization?
Checking, at the interval the case promised, whether the benefit actually arrived, and logging the answer beside the original estimate. It is fifteen minutes per project and the single highest leverage habit in investment decision quality.