Chapter 05
Building a defensible business case
A practical walkthrough of assembling a business case that survives finance scrutiny: structure, sources, NPV framing, sensitivity, and the rehearsal that surrounds the document.
01
Structure that the reader can navigate
A defensible case has a predictable structure. Executive summary that names the total as low — mid — high and breaks the mid by shape. Operational savings section with each line modelled. Risk mitigation section with each event identified, sized, and ranged. Capacity section with strategic argument and any defensible quantification. NPV summary with the discount rate stated and the assumption inputs visible. Sensitivity table covering the inputs that move the result most. Assumptions appendix. The structure does not need to be elegant; it needs to be navigable. A finance partner with twenty minutes should be able to find any single line and its source without help. The test is simple: hand the document to a reviewer who has not seen it before and ask them to find three specific assumptions in two minutes. If they cannot, the structure has failed regardless of how strong the analysis is underneath. A practical convention that pays back its small cost: number every line item across the case so the executive summary, the body, and the appendix all reference the same identifier. Line 4.2 in the executive summary points to line 4.2 in the operational savings section, which points to line 4.2 in the assumptions appendix. The numbering looks bureaucratic and reads as discipline. Reviewers can navigate by line number rather than by page. Authors can update a single line across three views without losing track. Cases written this way tend to age better than ones written without numbering, because future revisions remain navigable.
02
Sources, NPV framing, and the discipline of citation
Every quantitative claim cites its source. Internal time estimates from the team that does the work, dated. Cost rates from the budget partner, named. Industry benchmarks from named publications. Each source is referenced and re-checked at scheduled intervals. NPV framing matters when the contract life is more than a year. A typical RDI case runs 24 to 36 months, and discounting at the organisation´s WACC — often somewhere between 7 and 12 per cent — meaningfully shifts the result. A 60,000 euro saving in year three is not 60,000 euro today. State the discount rate. State whether the workflow cost is treated as opex or capex. State whether the benefits are pre-tax or post-tax. The conventions are boring; their absence is what kills cases. A case with hidden sources tends to fail at the first scrutiny. A case with visible sources, an explicit discount rate, and a stated tax treatment tends to pass even when the numbers are challenged. A worked NPV illustration. A 240,000 euro undiscounted three-year saving distributed roughly evenly across the period, discounted at 9 per cent, lands at around 202,000 euro in present-value terms. The 38,000 euro gap is not a flaw to be hidden; it is information the reviewer expects to see. State the gap, name the discount rate, and let the reader confirm the arithmetic. Cases that quote undiscounted totals to a finance audience tend to be challenged on first read regardless of the underlying quality. Cases that quote discounted totals tend to be discussed on the merits.
03
Sensitivity and the one-page table
A board pack that contains a one-page sensitivity table is unusually well prepared. Five to seven rows, each naming one input. Three columns, showing the total at the low, mid, and high value of that input. The reader sees in twenty seconds which inputs drive the answer and which are noise. A useful refinement is the two-way sensitivity for the two largest drivers — usually the proportion of report preparation time replaced and the probability of the largest risk event. A small grid showing the total at the low and high of those two inputs simultaneously communicates more than any narrative. As a rule of thumb, if the total at the worst corner of the two-way grid is still positive after workflow cost and discount, the case is robust. If it is negative, the case depends on those two inputs being right; the board should know that explicitly. A complementary artefact is the breakeven line: the value of the most sensitive input at which the discounted total goes to zero. If the breakeven proportion of report preparation replaced is 22 per cent, and the modelled mid is 55 per cent, the case has substantial headroom. If the breakeven is 48 per cent against a modelled mid of 55 per cent, the headroom is thin and the board should consider whether the modelled mid is genuinely defensible. Breakeven lines convert the sensitivity table from interesting to actionable. They turn a question of judgement into a question the reader can answer with their own intuition about the input.
04
The rehearsal: a business case is the conversation around the document
A business case is not a document; it is the conversation around the document. Walk the case through the finance partner before submission. Adjust where the partner spots weakness. Submit the version that has already survived a rehearsal. The discipline shortens the formal review and tends to produce the approval that the case earned in the rehearsal rather than in the meeting. The rehearsal has a shape. Send the case to the partner forty-eight hours before. Sit with them for thirty minutes. Ask three questions: which line do you trust least, which assumption would you challenge first, what would have to be true for you to be wrong about that. Take notes. Adjust the document. Submit. Most cases improve materially across the rehearsal. A few are revealed to be unready and quietly withdrawn, which is itself a win. The cost of the rehearsal is one hour. The cost of a rejected case is much higher than that. A complementary practice is the second rehearsal with the operational partner who will live with the workflow if the case is approved. They will identify weaknesses the finance partner cannot see — the legacy report someone senior insists on, the organisational dependency that has not been negotiated, the assumption that the affected team will reabsorb their saved time productively. The second rehearsal usually adds two or three caveats to the document and prevents the awkward post-approval moment in which an operational owner says they were never asked. Both rehearsals together cost less than two hours and substantially reduce the risk of approval-then-stall.
05
Board-pack defensibility and what reviewers really test
A board pack defensible under cross-examination has four properties. First, the total appears as a band, not a point, and the band is consistent with the line items. Second, every line cites a source the reviewer can verify in writing within ten minutes. Third, the sensitivity table identifies the inputs that matter and their effect. Fourth, the document acknowledges what would make the case fail. Reviewers test these in order. They check the total against the lines. They sample two sources. They look for a sensitivity table. They look for the failure case. A document that passes all four tests gets a focused conversation about substance. A document that fails any of them gets a procedural conversation that consumes the meeting. The framework is a public discipline; it is not specific to any vendor or product. A team that adopts it once tends to adopt it permanently because it makes the next case easier and the case after that easier still. The fingerprint of a strong document is therefore short and recognisable: a banded total, traceable line items, a one-page sensitivity table, and an explicit failure case. Reviewers who see all four know within five minutes that the rest of the document is worth their time. Reviewers who see three of the four ask a polite question to confirm the fourth. Reviewers who see two or fewer drift into a procedural conversation, and the substance never gets reached. A team that consistently produces all four signals tends to find that approvals shorten by ten or fifteen minutes per case, which compounds across a portfolio of cases per year.
Practice
01. Draft a one-page executive summary for a current case. It should contain the total as low — mid — high, a three-shape breakdown of the mid, the discount rate used, and a one-line statement of what would make the case fail.
Look for: A single page with a banded total, a three-line breakdown by shape, the WACC or hurdle rate used, and a sentence such as "the case fails if the proportion of report preparation replaced is below 25 per cent." Nothing else.
02. Schedule a thirty-minute rehearsal with a finance partner before your next submission. Write the three questions you will ask them. Capture their answers and note the two changes you will make to the document.
Look for: Three questions in writing — which line is least trusted, which assumption would be challenged first, what would have to be true for the partner to be wrong — and two named edits to the document arising from the rehearsal.
Checkpoint
For a current case, could you walk it through the finance partner before submission and show them a band, a sensitivity table, and the failure case in twenty minutes?
Recommended reading