Model Health

A one-look gauge of a model’s size, complexity, and obvious problems. Model Health scans the whole workbook and presents a single report — how many formulas, how complex, how many errors and where, which risk indicators are present, and how much of it is proof-covered — so you know what you’re dealing with before you dig in.

The scenario

Someone hands you a model and asks “is this any good?” Before reading a single formula you want the shape of it: how big, how complex, how many errors, how much is hardcoded, and whether anyone has vouched for the numbers. Model Health answers that in one report — a useful first stop at the start of a review, and a shareable snapshot of quality at a point in time.

Where to find it

ModelXcel Pro tab → Audit group → Model Health (keytip MH, screentip “Model Health Report”). It scans the active workbook and opens a report window (titled Model Summary). On a large model it shows a wait cursor while it reads every formula.

What the report contains

The report scans every worksheet and presents:

  • Counts — formulas and constants across the workbook, and a count of unique formula shapes (distinct R1C1 patterns).
  • Complexity — average and maximum complexity scores, and a distribution.
  • Errors — an error-cell count overall and per worksheet.
  • Risk indicators — error cells, hardcoded constants, merged cells, data validations, and the maximum complexity, as headline chips.
  • Top functions — the most-used functions in the model.
  • Per-worksheet grid — the same measures broken down sheet by sheet.
  • Proof coverage — a summary of the workbook’s proof-of-derivation state: how many inputs are bound, how many cells are signed, and whether the proof ledger’s chain is intact.

What changes in Excel

Nothing. Model Health only reads the active workbook and shows a report. The optional Export to Excel builds that report into a brand-new, unsaved workbook (a separate file you save yourself) — it does not write to or alter the model you analyzed. There is nothing to undo; close the report window to dismiss it.

A snapshot of size and risk — not a correctness check

The counts and risk indicators tell you how much is complex, hardcoded, or erroring — not whether the model is right. A model can look healthy and still be wrong. Treat Model Health as orientation, then verify the parts that matter with Proof & Trust, and drill into specific cells with Cell Search and the Risk Heatmap.

How it fits the review workflow

  • Start a review here for orientation, then use Cell Search to list the exact risky cells and Flag Cell the ones to follow up.
  • The proof-coverage figures mirror what Verify & coverage reports — a quick read on how much of the model has been signed off, which rises as you mark cells Verified.

Caveats & limitations

  • Read-only. The report changes nothing; the export is a separate new workbook you keep or discard.
  • It scans the whole workbook, so a large model takes a moment (a wait cursor shows while it works).
  • “Please open a workbook first.” — open a workbook before running the report.

Related