Finance does not fund efficiency. It funds a number attached to a decision. Here is how to build a case that survives being questioned.
Four kinds of value, in order of how credible they are
1. Time — credible only if you say where it goes.
Work it out honestly. Applications per year, times minutes per application, times the share you actually review today. That last multiplier is the honest part, and most business cases quietly leave it out to make the saving look bigger.
Then the crucial step: name what the freed hours are for. "Twelve hours a week of recruiter productivity" gets discounted to zero. "Nine hours a month, which lets one recruiter carry the two extra jobs planned for Q3 without hiring another recruiter" does not.
2. Cost — the strongest line.
Agency fees avoided is the best argument you have, because the alternative is a real invoice someone has paid before. One senior job filled in-house instead of through an agency often covers a year of tooling on its own.
Be conservative. Claim one job, not five. A modest number that holds up beats an ambitious one that gets picked apart — and the picking apart damages everything else in your case.
3. Quality — report it, do not model it.
Module 10 explained why this cannot be measured cleanly. Do not put a number on it in a business case. Report the stand-ins you track and say what they are: "we track whether hires are still here at 12 months, by source, as a rough quality measure."
Putting a made-up quality figure in a financial case is the fastest way to lose credibility for everything else on the page.
4. Risk — describe it, do not price it.
Being able to show your criteria, your records and your bias checks has real value that nobody can put a number on. Describe it in two sentences. Attempting to price it invites an argument you cannot win.
What not to claim
| Avoid | Why |
|---|---|
| Cost of a job being unfilled | Only credible if your business already models revenue per role. Most do not |
| A quality-of-hire improvement | You cannot measure the baseline, so you cannot show an improvement |
| Numbers from vendor case studies | They describe someone else's process. Everyone knows it |
| Time savings with no reallocation | Discounted to zero, and it makes your other numbers look optimistic too |
Show the downside before you are asked
Present two columns: your base case, and a version where every benefit is halved. If it still works in the second column, you have an argument. If it only works in the first, you have a hope.
Presenting the downside yourself does more for your credibility than any single number in the case.
Download the business case spreadsheet · all templates
The shape of a case that gets approved
One page. A specific saving attached to a decision someone was going to make anyway. A conservative agency-fee number. Quality named as a stand-in. Risk described in two sentences. And the downside case shown next to the base case.