01 · Ask
Start with the business question.
“What happens to runway if hiring slips a quarter?” belongs at the top of the workflow, not in a comment beside a spreadsheet tab.
Product
Finance analysis that keeps the receipts attached.
minrva is being built around the work FP&A teams already do: update assumptions, explain variances, trace drivers, and turn the answer into a concise operating narrative. The strongest product claim is simple: every answer should show its work.
02 · Inspect
Keep assumptions visible.
Every answer has to survive finance review: source lines, scenario changes, and a table tie-out below the visual.
03 · Explain
Write the variance narrative.
The output should help a CFO speak plainly: what moved, why it moved, and what decision it suggests.
Workbench views
Pick the question; the view follows.
Operators don't read finance reports top-to-bottom — they jump to the view that answers the question at hand. The workbench is organized the same way.
The answer reads like a research note.
Finance writes in figures, and figures carry citations. Each number points back to the ledger row, the budget version, or the assumption it came from — quietly, where finance expects to find them.
Q3 revenue came in at $1,180K1, $80K above plan2. The beat tracks to enterprise renewals (+$110K)3 and one early-stage expansion ($24K)4, partially offset by SMB churn (−$54K)5.
Sources
- GL.4000-Revenue 2026-Q3 · $1,180,043
- Budget.FY26-Q3 revenue plan · $1,100,000
- 2.aGL.4000-Revenue 2025-Q3 actuals · $1,000,000
- 2.bPlanning.assumption YoY growth · +10.0%
- =$1,000,000 × 1.10 = $1,100,000
- CRM.renewals stage=closed-won · 14 accts · +$110,228
- CRM.expansion acct=ACME-87 · MSA-2026-09 · $24,000
- Billings.churn 6 accts · YTD · −$54,302
Assumptions
- FX 2026-09-30 spot
- Rev rec ASC 606 · ratable
- Allocation by-customer-segment
verified 2026-10-14 09:32 PT · same inputs → same answer
The income statement, read left-to-right.
A Sankey flow shows revenue splitting into cost of revenue, operating expenses, and operating income — same data finance already reports, organized so the eye follows the dollar.
Income statement flow · synthetic SaaS company
Proportion audit: ECharts Sankey uses each link's numeric value directly; the $35M COGS band is 35% of the $100M revenue band, and Net Income is 12%.
Driver order should tell the story.
A bridge is not decoration. It should sort drivers into tailwinds and headwinds, label values inline, and end on the line item finance needs to explain.
Budget vs Actual · Q2 operating plan
Sorted by variance magnitude. Axis window $88M–$110M for homepage legibility.
| Step | Type | Δ $M | Running $M |
|---|---|---|---|
| Budget | Anchor | — | 100.0 |
| Revenue beat | Favorable | +8.0 | 108.0 |
| R&D under | Favorable | +1.0 | 109.0 |
| G&A higher | Unfavorable | −1.0 | 108.0 |
| S&M higher | Unfavorable | −2.0 | 106.0 |
| COGS higher | Unfavorable | −3.0 | 103.0 |
| Actual | Anchor | — | 103.0 |
Heatmaps should know revenue and expense rows differ.
A revenue miss and an expense underrun can both show negative numbers. minrva treats the accounting sign and the business signal as separate facts.
unfavorablefavorable
Story: revenue holds above plan, COGS pressure and heavier operating spend compress mid-year operating income, then Q4 recovery appears as costs come back under budget.
| Row type | Raw variance | Signal |
|---|---|---|
| Revenue | −5% | Unfavorable |
| Expense | −5% | Favorable |
Funded burn buys time; revenue closes the gap.
A revenue-versus-burn chart should show two honest curves and one decision moment: the month where revenue clears monthly costs. Everything before that is what the round actually pays for.
Story: investment burn funds the build during months 1 to 14; growth carries revenue past monthly costs around month 15. The shaded area is cumulative burn until break-even — what the round actually buys.
Actual against budget, with a forecast in the same frame.
A bullet graph compresses three numbers per line item — actual, budget, and forecast — onto one row. Favorability colors the bar so the eye finds risk before reading the label.
actual (favorable) actual (unfavorable) budget forecast (full-year)
Read it: each bar is YTD actual against a YTD budget line; the dashed mark is the full-year forecast. Bars colored by favorability — for expense rows, under budget is favorable.
Pilot program
Bring the messy workflow. We will test against that.
The first pilots should have real FP&A work to pressure-test: board prep, cash planning, variance meetings, or scenario updates that currently take too many exports and too much manual explanation.