Sign-off at 5pm Friday
A consultant has prepared a feasibility report that a client will lodge with a government department on Monday as part of a grant application worth $2.4 million. The report's central case assumes revenue growth of 18% per annum for five years. The assumption came from a spreadsheet the client supplied with no supporting evidence; the analyst who built the model left the firm three weeks ago. A second-year consultant inherited the file and cannot explain how the 18% was derived. The engagement partner wants sign-off by 5pm Friday because the client's board meets on Monday morning and the lodgement deadline is fixed by the programme. It is 2pm. The reviewer has roughly three hours available and two other files to finish that day.
1. What is actually at stake in signing this report, and what are the realistic options?
Read the worked answer
What sign-off means. A signature on a lodgement document is a representation that the work was performed competently and that the conclusions are supported. It is not a statement that the client supplied the number. The department will make a $2.4 million decision relying on it, and the firm's name is on it.
Why the 18% is the whole problem. It is not a small input - at five years it compounds to roughly 2.3 times revenue. Every downstream conclusion (capacity, staffing, premises, debt) rests on it. An assumption that large, personally unexplained, with no source, is not an assumption; it is a gap.
The realistic options:
- Sign as-is. Fastest, and the only option that cannot be defended if challenged.
- Refuse to sign. Protects the firm, strands the client, and may be an overreaction if the number turns out to be defensible.
- Sign a report that is honest about the basis - source the assumption or replace it with a defensible base case, and present the client's 18% as an upside scenario with its provenance stated.
Option 3 is almost always available and is the only one that serves both the client's deadline and the firm's position.
2. You have three hours. What is the minimum acceptable standard, and how do you get there?
Read the worked answer
Three hours is enough for the thing that matters and not enough for everything.
Do first (the 18%):
- Go back to the client's CFO today, in writing: where did 18% come from, and can they give the basis? A published industry forecast, a historical average, a board target - any of these is a starting point. A board target must be labelled a target, not a forecast.
- Run the model at a defensible base case (say, historical growth or a sector benchmark) alongside the client's 18% case. Present both. A report that shows the outcome under two growth assumptions is more credible than one that asserts either.
- Include a sensitivity: state the growth rate at which the project still meets the grant criteria. That single number is what the department will most want to know and is quick to compute.
Do next: a limitations section that names the assumptions the firm did not independently verify and their source. Specific beats generic.
Consciously do not do: rebuild the entire model. The task is to make the report defensible, not perfect, and three hours spent rebuilding leaves no time for the assumption that drives everything.
3. At 4:30pm the partner says: 'The client accepts the 18%. Just sign it and note it's client-supplied.' What do you do?
Read the worked answer
Client acceptance is not the test. The test is whether the report, read by the department, fairly represents the basis of the projection.
What that instruction would produce: a sign-off with a note that the key driver is unverified. That is honest only if the note is prominent and the report shows what happens if the number is wrong. If the 18% stands alone as the base case with a footnote, the note is doing decorative work.
The response:
- Offer the version that achieves the partner's goal - lodgement on Monday - without the exposure: the dual-scenario report with a named source for each. It takes the remaining time and no more.
- Put the discussion in writing. Not as an escalation threat, but because the partner's decision may be reasonable on facts the reviewer does not have, and the record protects both people whichever way it goes.
- If instructed to sign a report that asserts an unverified assumption as the base case with no sensitivity, decline to sign personally and ask that the instruction be recorded. A signature is personal.
The reasoning to hold onto: the deadline is the client's risk, not the firm's. Missing Monday costs a grant application. Signing something unsupportable risks the firm's standing with the department on every future lodgement - and the department will remember which firm signed.