Illustrative sample · synthetic company · not a client
Every name and number below is invented. This is three findings from an audit deliverable, written at the depth a paid one is, though never the whole of it. The audit also produces a stakeholder and workflow map, a data and integration assessment, a prior-art review, a risk register and a ranked shortlist. Judge the thinking on the part we show.
Decision Opportunity Audit · sample deliverable
Northwind Digital: which projects are quietly losing money?
A 64-engineer software consultancy, ~€14m revenue, eleven active engagements across fixed-price, capped and milestone contracts. The owner's question: are we as profitable as the P&L suggests, and if not, where, and what can we still change? Below is the two-to-three-week audit's output, compressed to one page.
Verdict
Reshape, don't stop.
3 of 6 sampled projects are heading below zero margin, 2 of them in one account (Retail A), and all 3 on the same contract type (Fixed price). All are still recoverable this quarter.
Margin at risk
≈ €1.17m
How far under water the sampled portfolio runs at p90: the sum, across projects, of forecast cost above contract value. 68% of it sits in two projects with the same root cause.
Root cause
Unpriced scope + review queue.
Not slow engineers. Scope grew without a change order, and a two-person review bottleneck turned rework into calendar time.
Finding 1 · the portfolio, forecast to completion
Cost at completion, per project
Each project's remaining cost is forecast as a distribution. The p90 column is the bad-but-plausible outcome. The probability of loss is the mass of that distribution above the contract value.
| Project | Client | Contract | Value | % done | Cost p50 | Cost p90 | P(loss) |
|---|---|---|---|---|---|---|---|
| P-01 | Retail A | Fixed price | €1.85m | 72% | €1.98m | €2.31m | 68% |
| P-02 | Bank B | Capped T&M | €2.60m | 40% | €2.35m | €2.72m | 22% |
| P-03 | Retail A | Fixed price | €0.90m | 55% | €1.03m | €1.24m | 74% |
| P-04 | Logistics C | Milestone | €3.10m | 30% | €2.55m | €2.98m | 9% |
| P-05 | Health D | Retainer | €1.20m | 60% | €1.05m | €1.22m | 14% |
| P-06 | Bank B | Fixed price | €1.40m | 85% | €1.44m | €1.63m | 55% |
| Sampled total | €11.05m | €10.40m | 3 of 6 at risk | ||||
Shaded rows: probability of finishing below zero margin ≥ 50%. Method: lognormal cost-to-complete forecasting, calibrated on the client's own historical deliveries.
Finding 2 · where the margin goes
Where the margin goes
Decomposing the two worst projects (P-01, P-03, the same client) shows the money leaks in four places, and "the team is slow" is not one of them:
- Unpriced scope growth ≈ 41%
Nine change requests delivered, one change order raised. The work shipped; the invoice did not.
- Review-queue rework ≈ 28%
Two senior reviewers gate every merge. Median time-in-review is 3.1 days. Rework on already-merged code is high.
- Staffing mix ≈ 19%
Two specialists on a project that a mid-level could carry for 60% of it, because the plan never freed them.
- Customer-created delay ≈ 12%
Sign-off latency the contract does not charge for. The smallest of the four levers, and a contract problem rather than an execution one.
Finding 3 · what can still be changed
Four moves, ranked by the margin they recover
- 1
Renegotiate P-01 / P-03 scope now
The nine change requests are documented. A single consolidated change order, raised this week, is worth more than any engineering change, and the window closes at delivery.
- 2
Add a third reviewer, time-box review
The bottleneck is two people. The tooling is fine. A third reviewer and a 24-hour review SLA cuts the rework tail on every project, and these two most.
- 3
Re-mix P-01 staffing
Free one specialist to a project that needs them. Back-fill with a mid-level. Protects margin here and delivery elsewhere.
- 4
Stop bidding this contract shape
Fixed-price for this client, at this scope-volatility, has lost money twice. Price the volatility in, or decline. Do not run it a third time on hope.
What this audit does not claim
- That the forecasts are certain. They are distributions, and the report states the coverage of its own intervals against the client's history.
- That the effort estimates are right. The audit tests how sensitive the verdict is to them, and says which findings survive a pessimistic estimate.
- That software is the answer, for Northwind, the first three moves are contractual and organisational. A system to watch this continuously is a separate, later decision.
An audit is useful even when its recommendation is "change the contract, not the code." That is the output here, and on these numbers it is worth about forty times what the audit costs.
Start with the decision
The same audit on your own portfolio
This was a synthetic company. On your own delivery data, with your baselines and your contracts, the same audit tells you which of your projects is drifting, and what you can still do about it. That is a Decision Opportunity Audit, from €15,000.