Skip to main content
Numbers, honestly

How to read every figure on this site.

We put numbers everywhere because CFOs check them. This page is the legend — so a careful reader never has to guess what “recoverable” means.

1. Illustrative vs measured

Homepage examples, labs, and the ROI calculator are illustrative — built from our method so you can size the opportunity. They are not “typical results from Snapshots we've run” as a performance claim. Your free Snapshot replaces every illustrative figure with a measured one from your own work.

2. Capacity is not cash

Most “$/yr” figures are capacity valued at loaded labor cost (hours × fully loaded salary). That is the cost of the work today. It becomes P&L cash only if you avoid a hire, redeploy people to higher-value work, or catch real leakage (duplicates, missed credits, lost discounts). We assume no layoffs unless you say otherwise. Longer take: Capacity is not cash.

3. Two ways we size the opportunity

Team ballpark: about 3% of your whole team's loaded labor. Broad and deliberately low. At ~$95K base, 45 people land near≈$166K/yr, 220 near ≈$815K/yr, and 500 near≈$1.85M/yr. The AI recovery calculator uses this band.

Process estimate: 55–70% of the hours inside one targeted process, such as the close. That is why the Close Cycle Lab and the Sample Audit show a larger number than the team calculator: they look at a narrower, heavier slice. AP Leak and ARR Bridge are leakage models, not labor models. Working Capital and PE Hold are separate cash theses.

Both layers are illustrative until your Snapshot measures them. They are not one calibration, and they are not meant to match.

4. What CFOs and PE partners can bank

  • Days off the close — calendar and team capacity
  • Hires avoided as you grow — FTE you do not add
  • Cash collected sooner — decision lag and DSO/CCC timing
  • Leakage caught — duplicates, overpays, missed credits
  • Hours returned to analysis — not “headcount cut”

5. Labs use fixed seeds

Monte Carlo labs use a fixed random seed so P50s do not drift between refreshes. Move a slider and you get a new distribution on purpose.

6. Plain-English glossary

  • Loaded labor — base salary × ~1.3 for benefits, taxes, and overhead. The cost of an hour of your team's time.
  • FTE — full-time equivalent. ~1.0 FTE ≈ one person's capacity; usually absorbed as growth, not a layoff.
  • P10 / P50 / P90 — percentile outcomes from a Monte Carlo run. P50 is the middle case. For close days, P10 is the shorter tail and P90 is the longer tail. For dollar outputs, P10 is the lower dollar tail and P90 is the higher one. Read the direction from the lab.
  • People who do this work — the team on the repetitive process, not total company headcount.
  • Claude — Anthropic's AI model. We build on it inside your accounts; you carry the ongoing license.
FAQ

Common questions

Are lab numbers promises?

No. Labs and homepage examples are illustrative capacity bands built from our method. The free Snapshot replaces every figure with a measured one from your own work.

Is recoverable dollars the same as cash?

Usually not. Most $/yr figures are capacity valued at loaded labor cost. They become P&L cash only if you avoid a hire, redeploy people, or catch real leakage. We assume no layoffs unless you say otherwise.

Why do Monte Carlo labs repeat the same P50?

Labs use a fixed random seed so P50s do not drift between refreshes. Move a slider and you get a new distribution on purpose.

Related paths
Next step

Replace every figure with yours.

Labs and examples are illustrative. The free Snapshot measures the same story on your own work — yours to keep.