What every outcome record contains
We record each engagement's results in the same five parts, so outcomes can be compared and challenged.
| Part | What it captures |
|---|---|
| Baseline | Effort, quality, cycle time and error rate before any change, measured on real work for long enough to be representative. |
| Intervention | Exactly what changed: tool, configuration, process step, review step, training. Including what we chose not to do. |
| Measured outcome | The same measures after the change, over a comparable period, with the size of the sample. |
| Limitations | What the numbers don't show: small samples, seasonal effects, novelty effects, self-reported time. |
| Maintenance burden | Ongoing review time, licence and run costs, evaluation re-runs and the owner responsible. |
How we label claims
Every material statement in a brief, assessment or case study carries a source, a date and one of these labels. The labels stop weak signals from being turned into made-up demand, and stop demos from being mistaken for results.
- Observed: Seen directly in public material or the client's systems, with a source and a date.
- Inferred: Our reasoning from observed facts. Never presented as fact, and never used to claim a need or a budget exists.
- Buyer-confirmed: Stated by the accountable person in the organisation.
- Tested: Measured by us on the client's own examples, against a threshold agreed in advance.
- Expired: Was true once, but the model, vendor, process or team has changed since. Has to be checked again.
Vendor claims are not test results
A vendor's benchmark tells you what a product can do under its maker's conditions. A test result tells you what it did on your work, judged against your threshold. We report the two separately and never combine them into one figure.
Hours released are not cash saved
If a workflow redesign frees up staff time, that capacity has value only once it's used. A business case has to say what happens to it: less overtime, a hire avoided, more throughput, better service — or nothing. Running costs come off the total: review time, licences, support and change management. The workflow economics calculator keeps these figures separate.
Worked examples (hypothetical)
These examples are illustrative. They are not client results. They show the method, not a benchmark.
Example A: 120 staff-hours a month
A redesigned reporting workflow releases an estimated 120 staff-hours a month at a loaded cost of £45 an hour. That time is worth £5,400 a month, but it is not cash.
- If the team was paying for 60 of those hours as overtime, about £2,700 a month becomes a real saving.
- If the rest goes on client reviews, the benefit is service quality. That's real, but it belongs in the business case as quality, not cash.
- Take off £600 a month for licences and review time, and £8,000 one-off for set-up and change management. The cash case pays back in about 4 months, and only if the overtime really stops.
Example B: the margin on a £20,000 implementation
For a principal-delivery engagement, the gap between the customer fee and a specialist's quote is not profit.
| Item | Assumed amount |
|---|---|
| Customer fee | £20,000 |
| Specialist delivery | −£9,000 |
| Diagnosis, oversight and acceptance work | −£3,000 |
| Tooling and rework allowance | −£1,000 |
| Acquisition and presales | −£2,000 |
| Contribution before overhead and tax | £5,000 |
We publish this example so buyers can see what they are paying for: the delivery, plus the accountability for whether it is accepted.
Published case studies
None yet. We publish case studies only with the client's written permission, with the five-part record above, and with their limitations stated. Until then we would rather show our method than make up a track record.
What we won't do
- Present one engagement as an industry-wide benchmark.
- Publish a client's confidential information, or use it to write public content.
- Rank suppliers, or present a recommendation we have a commercial interest in as independent.