Larkspan Software: what a buyer will find
A made-up 85-person SaaS, to show the format of the 6-week baseline. Every number in yours comes from a script your team can re-run.
Other samples: Product team diagnostic · Operations diagnostic · Pre-sale R&D baseline
R&D % of revenue31%peers 20–28% · run 58% · build 42%
Capitalisable, candidate36%of R&D hours on new features; your auditor decides
Lead time / CFR11 d · 14%peers 4–9 d · 10–15%
Planned vs delivered64%peers 70–85% · last 4 quarters
Where a published benchmark fits your size, the report names it beside the range; where none does, we say so and show no range.
Key-person risk
In blue, systems where one person does over 80% of commits and reviews. Billing and Reporting are what a buyer's diligence team will ask about first.
Tech-debt hotspots
| Area | Why | Est. cost to fix |
|---|---|---|
| Billing proration | changed in 40% of releases, 9 incidents in 12 months; cost = estimated engineer-weeks × loaded weekly cost | €85k |
| Report generation | slowest endpoint, single maintainer | €40k |
| Legacy admin | no tests, rarely changed | leave |
Licences, security, running cost
| Check | Finding | Result |
|---|---|---|
| Open-source licences | 412 packages; one AGPL library in the reporting service | Replace |
| Known vulnerabilities | 0 critical, 3 high (all in one outdated PDF library) | Fix pre-sale |
| Hosting and third-party cost | €41k / month; 6.1% of revenue (peers 5–9%), 8.4% of gross margin | In range |
AI claims check
| Claim in the deck | What runs | Verdict |
|---|---|---|
| "AI-powered forecasting" | gradient-boosted model, retrained monthly, used by 70% of accounts | Verified |
| "AI assistant for every user" | prototype behind a feature flag, 3 pilot customers | Overstated |
| "Automated anomaly detection" | fixed thresholds, no model | Reword |
A confidential call about your timeline.
A rough sale date is enough to start. On the call we'll tell you which of a buyer's questions your current data can already answer, and which it can't.