Understanding the Review Report

What is the Review Report?

Every model gets an automated review - the Review tab - that checks whether your assumptions make sense. It runs deterministic checks (benchmark comparisons, consistency rules, impossible-input detection, and business-model-specific logic for your vertical) and summarizes the result as a Health Score from 0 to 100. Think of it as a seasoned CFO reviewing your numbers before an investor does.

It works for every business type, with checks tuned to each vertical. The free tier gets limited insights; the full report comes with Build It Yourself.

---

Whose Numbers Are These?

The report starts with provenance - how much of the model is yours:

The "Whose numbers are these?" meter ("You've personalised 1 of 28 tracked assumptions") and "Defaults a VC will challenge first" - the high-impact assumptions still at their benchmark default.

• The personalisation meter splits tracked assumptions into Your data vs Benchmark defaults
• "Defaults a VC will challenge first" lists the high-impact assumptions still at their industry default - "review them before you pitch"

Assumption-to-Refine Tags

Wherever a value came from an industry default rather than from you, the model tags it as an assumption to refine:

On the defensibility card, default-driven strengths are labelled "assumption to refine" - e.g. "Non-payroll S&M as % of Revenue: assumption to refine". The tag clears once you replace the default with your own number.

Defaults are good starting points - and honest labels. Investors ask "where did this number come from?", and the tags make sure you know which answers are still "an industry benchmark" rather than "our data".

---

The Health Score and the Formula Explorer

Severity Levels

• Critical - plan-level problems, e.g. "Held funds coverage shortfall. In month 58, cash on hand ($291K) is below the seller funds held in trust ($382K)... The business would be financing itself with money owed to sellers."
• Warning - benchmark deviations, e.g. "Monthly GMV Growth Rate Below Benchmark - your value of 3.4% falls below the typical benchmark range (5.0% - 20.0%). Investors will likely flag this."
• Info / opportunities - observations and quick wins

Tracing a Number to Its Drivers

Click any formula output in the list and the right-hand panel shows the assumptions that drive it, editable inline - "Select a formula output to see and edit the assumptions that drive it." Fix the driver, and the flag clears on recompute. The search box and Sheet / Severity filters help in bigger models, and the trace hints point you at the single assumption that clears the most flags.

---

A Suggested Workflow

1. Fix criticals first - they're credibility-enders in diligence
2. Work the "defaults a VC will challenge" list - replace each with your real number (or consciously keep the benchmark, knowing you'll be asked)
3. Then judge the warnings - deviation from benchmark isn't automatically wrong; being unaware of it is
4. Re-run your eyes over the score - most models can't reach 100, and that's fine. The goal is knowing exactly why you sit where you sit.

Tip: The planning preview's defensibility card ("X things an investor will challenge" + Make it defensible) is the same engine in summary form - the button drops you straight into this tab.