For Corporate FP&A & Accounting
Forecast variance? You'll know the cause before the meeting.
Group FP&A · controllership · subsidiary controllers · corporate accounting
The one capability that lands you
Why This Number
Click any cell, see the full causal chain — sources, formulas, signatures, and timestamps.
Why This Number
Click any cell, see the full causal chain — sources, formulas, signatures, and timestamps.
Demo comingWhat changes for you
Three concrete shifts.
- 01
Why This Number answers the CFO's "why is this different from last week" without a manual hunt.
- 02
Compare to Previous Version with Material Change Intelligence pinpoints exactly which forecast driver moved.
- 03
Audit Trail keeps a persistent record across the close cycle — the .mxaudit sidecar travels with the workbook.
Supporting capabilities
Other features this persona uses regularly.
Same product, same install. The 7 features below are the ones Corporate FP&A buyers tell us they reach for after the lead capability.
Compare to Previous Version
Diff the active workbook against any earlier file on disk. Auto-discovery walks the workbook's folder for sibling files matching common version-suffix patterns (`*_v*`, `*_version*`, `*_rev*`, `*_backup*`); Browse fallback for anything else. Cloud workbooks (OneDrive, SharePoint) auto-resolve to the local synced copy or get snapshotted to a cache folder so the version picker has a folder to enumerate. Once selected, the same Compare Workbooks engine runs and "Explain Output Change" routes into Material Change Intelligence from the results window in one click.
Material Change Intelligence
Answer "why did Project IRR move from 12.4% to 11.8%?" with attribution math instead of guesswork. Auto-detects the model's outputs across the standard PF/IB metric vocabulary (IRR, NPV, DSCR, LLCR, PLCR, MOIC, Payback, Equity Return) via named-range and label-context scans; a type-ahead lets you specify any cell. Decomposes the delta into up to 8 ranked drivers. Sensitivity-backed contributions actually re-perturb the live workbook ±1% via the Sensitivity engine — a real local Jacobian, not pattern matching. Residual is exact arithmetic: Δ minus Σ contributions. Confidence blends evidence type, path length, and unexplained residual into a single score per driver. For each driver: a cell-by-cell Dependency Path showing the chain from the changed input through every formula it touched on the way to the output. Seven one-click routes from the result — Open In Excel, TrustLens, Formula Diff, Worksheet Map, Create Review Issue, Mark For Audit, Export Report. And the closed loop: every comparison that intersects a signed block writes a BlockBroken record to the Proof ledger; every explanation has an ExplanationId that links Review Issues and the Audit Trail.
Replay Proof
Pick any output cell that has changed. Replay Proof saves a copy of the baseline workbook to temp, opens it in live Excel, applies each revised driver, and recalculates the output after each step — actual formulas, UDFs, and named ranges executing in Excel, not a simulator. Up to 8 ranked drivers; per-driver dependency path from input to output; per-driver trust status (vector fresh/broken, block signed/broken, bound, no record). Two layers of integrity: every export is a tamper-evident SHA-256 self-hash certificate, and `.mxreplay` exports also write a sibling `.mxreplay.signed` envelope — a real Ed25519 asymmetric signature (BouncyCastle Ed25519Signer) over the canonical certificate plus envelope metadata, with the public key embedded in the envelope. Per-user keypair generated on first launch, private key sealed with DPAPI at `%APPDATA%\ModelxcelPro\keys\`. The repo ships a zero-dependency Node verifier (`tools/verify-replay/verify-replay.js`) that any auditor can run against `.mxreplay.signed` without installing ModelxcelPro. Issuance writes a CertificateIssued record back into the workbook ledger so the audit trail captures every export.
Audit Trail
Mark cells Verified, Questionable, or Erroneous. Status persists in a .mxaudit JSON sidecar that lives next to the workbook on disk and auto-saves on every change — send the file plus its sidecar to your auditor and they open the model exactly as you left it. Status changes also feed the Proof ledger; bulk edits that revoke verified status are recorded as SignatureRevoked records so the loss is auditable after the fact.
Sensitivity AnalysisCtrl+Shift+S
Multi-output sensitivity in one form. Auto-detects the model's outputs (≥0.9 confidence joins via the same TopologyDiscoveryService MCI uses) — switch between IRR / NPV / DSCR / LLCR / PLCR / MOIC / Payback / Equity Return without rebuilding. Up to 100 drivers per output, each perturbed at six levels (±1% / ±5% / ±10%) — not just ±1% — to show non-linearity. A Quick-Scan pre-pass with a 0.5%-impact diminishing-returns cutoff trims dead drivers before the full run. UI: Top-N filter (10/20/50/All), hide-zero, hide-outliers, click-bar to navigate to the input cell, "Copy Data" export, ETA-aware progress on a background task so Excel stays responsive.
Model Health Report
Workbook-wide diagnostics on one page: formula counts, complexity distribution (so you know which sheets are doing the heavy lifting), error breakdown, hardcoded-constant count, volatile-function flags, and risk indicators. The one-page health check you'd otherwise build by hand with COUNTIF and INDIRECT.
Direct style shortcutsAlt+Shift+N/P/D/F/R
Skip the classifier when you already know the answer. Alt+Shift + N / P / D / F / R apply Number, Percent, Date, Factor, or Ratio directly to the selection. Inputs auto-route to the input variant so the look stays consistent.
Other solutions