Financial modeling
Models that keep up with the business. Live data, your logic.
Cube is the financial modeling software that keeps your models on live actuals, with your formulas, drivers, and dimensions in one clean, decision-ready layer.
The problem
A model you never stop rebuilding.
The file is v27_FINAL, the links are broken, and the meeting is in an hour.
Actuals arrive by hand
Every cycle opens with exports, paste-overs, and a formula audit before the first real question gets asked.
Fragile by design
One renamed tab or broken link and the whole file is suspect until someone traces the damage.
One scenario at a time
By the time the what-if is rebuilt in a copy of the file, the question has already changed.
Modeling, the manual way
Exports, paste-overs, and a formula audit before the first real question gets asked.
The version-control system the modeling cycle actually runs on.
The distance between "the links are broken" and the meeting.
What replaces the rebuild: live actuals, drivers finance owns, and scenarios on demand.
Under the hood
How financial modeling software should work.
Cube delivers trusted, decision-ready data, everywhere you work, and where AI can do its best work.
The Agentic Finance Layer
The governed foundation powering every experience.
- Unified DataOne source of truth for all your business data
- Advanced Business LogicRules, metrics, calculations and consolidations
- AI Agents & AutomationInsights, forecasting, workflows and automation
- Security & ControlsAccess, auditing and compliance
- APIs & IntegrationsOpen APIs and pre-built connectors
One connection. Trusted everywhere.
The moment you need a consultant to change your model, finance lost control of the process.
Bi-directional sync
Build the model once. It stays live.
Fetch actuals straight from the clean layer, build the plan around them, and publish it back. The model you open in April already knows what happened in March.
| $K | Jan · Actuals | Feb · Actuals | Mar · Actuals | Apr · Plan | May · Plan |
|---|---|---|---|---|---|
| SaaS revenue | 1,860 | 1,905 | 1,948 | 1,990 | 2,035 |
| Services revenue | 320 | 310 | 336 | 330 | 335 |
| Revenue | 2,180 | 2,215 | 2,284 | 2,320 | 2,370 |
| Headcount (driver) | 142 | 145 | 148 | 152 | 155 |
| Operating expenses | 1,684 | 1,712 | 1,741 | 1,790 | 1,822 |
| EBITDA | 496 | 503 | 543 | 530 | 548 |
Charted
The model, drawn.
Revenue and EBITDA straight from the model. Actuals in navy, plan in gray, and every point traces to the source.
Dimensions
Change the rollup and keep the model.
Map every GL account once. When the company reorgs, teams move and departments split, and every number still reconciles to the same source.
- Operating expenses
- GTM new group, FY27
- Sales 6000, 6100
- Customer Success moved · 6300
- R&D
- Engineering 6000, 6200
- Product split · 6000
- GTM new group, FY27
Reorg the tree, keep the model. Same GL accounts underneath, so FY26 and FY27 still reconcile line for line.
Why teams choose Cube
One model that is always ready.
Actuals arrive on their own
Connect NetSuite, Salesforce, and Workday once. Every model opens on numbers that already tie to the GL.
Your logic, in plain formulas
No proprietary syntax and no consultant queue. Finance changes drivers, dimensions, and rules directly.
Numbers you can defend
Any figure in any model traces back to the source transaction, so the answer to "is this right?" is yes.
The model is already current
Actuals arrive on their own, so the model you open in April already knows what happened in March.
Adoption
Easy to start, hard to outgrow.
Cube deploys alongside the stack you run today, connects to hundreds of source systems, and stays in your team's hands.
Finance-led setup
Configured by your team, guided by ours.
Of sources
Pre-built connectors for major platforms, flexible methods for the rest.
Rip-and-replace
Your spreadsheets and your stack stay in place.
Onboarding
Hands-on setup with a named contact.
Finance-led from day one
Your team changes drivers, dimensions, and rules directly, guided by ours. No code, no consultant queue, no rip-and-replace.
Cube in action
FP&Agents keep the model moving while you sleep.
Purpose-built agents run on your clean, decision-ready data. Every output traces back to the source.
The Data Manager
Keeps every source mapped, reconciled, and current, so the model never waits on an export.
The Analyst
Explains variance to plan and drills to the transactions behind it before the meeting starts.
The Forecasting Agent
Smart Forecasting builds predictions from your history and seasonality, in your model's own drivers.
Q3 sales hires: 4
Q3 sales hires: 8
Q3 sales hires: 12
One driver changed. Everything downstream recalculates.
The model doesn't keep your hours
Agents keep sources mapped, explain variance, and build the forecast overnight. Every output traces back to the source.
Where you work
One model, on every surface.
Cube delivers the same governed numbers wherever the work happens. Explore the full story on Where You Work.
Cube Workspace
The browser app where finance builds the model, sets permissions, and publishes.
Excel & Google Sheets
Bi-directional sync: fetch live actuals into the sheet, publish plans back.
AI assistants via MCP
Claude, ChatGPT, and Gemini answer from your governed numbers.
Slack & Teams
Ask a question in chat and get the governed figure back, with the trace.
PowerPoint & Google Slides
Decks with figures bound to Cube that refresh to the current numbers.
BI tools
Tableau, Looker, and Power BI read from the same model as the plan.
Use cases
Where teams put the model to work.
Revenue modeling
Cohort models, ARR waterfalls, and renewal forecasts on live pipeline and billing data.
Headcount planning
People plans and financial targets in the same model, synced with your HRIS.
Scenario planning
Flex a driver and answer the what-if before the meeting ends.
Rolling forecasts
Reforecast as actuals land instead of once a quarter under duress.
Cash flow forecasting
A rolling view of cash position built on the same clean layer.
Budget vs. actuals
Variance explained, with every delta traceable to the transactions behind it.
Your workflow is on this list.
Pick the one that hurts most. The demo starts there.
Security & governance
Controls your auditor will recognize.
The model holds numbers that reach the board, so the controls are part of the product. Details live in the security overview.
Role-based access
Enforced at the cell level. The CFO sees everything; a department head sees their cost center.
Full audit trail
Every change to logic, mappings, and data, time-stamped and attributable.
Versions and approvals
Plans, forecasts, and scenarios kept as named versions with approval workflows.
Trace to the source
Every figure in every model maps back to the GL transaction behind it.
Trace to the source
Every number opens.
Every finance leader knows the feeling of presenting a number they can't fully explain. In Cube, the figure on the slide opens to the accounts beneath it and the transactions beneath those.
| Mar 03 | Snowflake | annual true-up | NetSuite | $38,400 |
| Mar 11 | Figma | seat renewal | NetSuite | $12,750 |
| Mar 27 | Datadog | usage overage | NetSuite | $9,310 |
| + 41 more transactions | ||||
Flex a driver and answer the what-if before the meeting ends.The live model · Cube
Already convinced?
Book a demo and skip the homework. Still researching financial modeling? The rest of this page is the long answer.
Definition
What is financial modeling software?
Financial modeling software
Financial modeling software is a platform for building and running the models a finance team plans the business with: revenue, headcount, expenses, and cash. It connects to source systems so actuals flow in automatically, keeps formulas, drivers, and dimensions in one clean, decision-ready layer, and delivers the outputs wherever decisions are made, from spreadsheets to board decks to AI assistants.
That layer is what Cube calls the Agentic Finance Layer: the model lives in one place, and every surface reads from it.
Actuals and assumptions in
Decisions out
The hard parts
Why is financial modeling difficult?
Not because the math is hard. Because the inputs move, the logic hides, and the org keeps changing underneath the file.
Data assembly eats the calendar
Most of the cycle goes to collecting and cleaning inputs. Connected sources remove that step entirely.
Formulas break silently
One renamed tab and the output is wrong with no warning. In Cube, logic lives in the layer and survives the file.
Versions multiply
Budget v3, forecast v7, the board copy. Cube keeps versions and scenarios inside one model.
The org keeps changing
Reorgs break rollups built into cell references. Dimensions remap once and every report follows.
What-ifs take days
A real scenario used to mean rebuilding half the file. Drivers in Cube recalculate everything downstream.
Nobody fully trusts the output
A model no one can audit is a model no one defends. Trace to the source ends that argument.
Before and after
What changes when the model moves off the desktop?
| Spreadsheet-only | On Cube | |
|---|---|---|
| Actuals | ✕Exported, pasted, and re-checked each cycle | ✓Fetched live from connected sources |
| Logic | ✕Hidden in cell formulas with one owner | ✓Centralized, readable, and changed by finance |
| Scenarios | ✕A copy of the file per case | ✓Drivers flex in place and every case recalculates |
| Versions | ✕v27_FINAL_final | ✓Named versions with approvals and history |
| Trust | ✕"Let me get back to you" | ✓Any figure traces to the transaction |
AI can be 80% right. In finance, that's 100% wrong.
A field guide
What are the main types of financial models?
Three-statement model
P&L, balance sheet, and cash flow linked so a change in one flows through the others.
Driver-based model
A handful of operational drivers (reps, pipeline, pricing) generate the line items.
Scenario model
Downside, base, and upside cases built from the same logic with different assumptions.
Headcount model
Hiring plans, comp, and loaded costs tied to the financial plan they land in.
Revenue model
Cohorts, ARR waterfalls, and renewals showing what drives growth and what is at risk.
Cash flow model
A rolling view of cash position, because revenue means little if you can't fund next quarter.
Vendor-neutral
Financial modeling best practices
These hold whatever tool you use. More depth on each lives on the Cube blog.
Separate inputs, logic, and outputs
An assumption should never hide inside a formula.
Model drivers, then let line items follow
Plan the causes and calculate the effects.
Keep one source of truth for actuals
Every model reads the same numbers, or the meeting becomes a reconciliation.
Version deliberately
Snapshot before each planning cycle so history stays comparable.
Build dimensions that mirror the business
Rollups should match how the company is managed, and survive the next reorg.
The checklist
What to look for in financial modeling software
Direct connections to your sources
Hundreds of source systems, with pre-built connectors for the major platforms you run.
Spreadsheet-native, with bi-directional sync
Fetch actuals into Excel or Google Sheets and publish plans back.
No-code logic finance can own
If changing a driver requires a consultant, the tool owns your process.
Dimensions, versions, and scenarios in the core
Structure belongs in the platform, never bolted on through file copies.
Traceability on every figure
Any number in any output should open to the transactions behind it.
AI that runs on clean, decision-ready data
Agents are only as trustworthy as the layer they read from.
Your spreadsheets can stay. Everything underneath gets smarter.
Models that keep up with the business.
Watch a live model fetch actuals, flex a scenario, and trace a number to its source.
Book a demoRuns on sandbox data, scoped to your stack and your questions.
