Skip to main content
Sigma for forecasts

Own every forecast version

Finance and sales ops set the drivers, run the scenario, and stand behind every version when the board asks.

Request demo
What the forecast cycle costs today

What the current cycle costs

Each number below comes from finance teams describing the cycle they run today.

A cash forecast lives in 18 tabsone workbook, needs 8 to 15 rebuildszero rebuilds a year, and costs a full daya few seconds per new column.

How it works

3 steps to a live forecast

  1. 1
    Connect the warehouse

    Sigma queries live warehouse tables, so the forecast always reads current actuals without an extract to refresh.

  2. 2
    Set the drivers

    Assumptions live in Input Tables next to the data. Every edit writes back to the warehouse with the user and timestamp.

  3. 3
    Freeze and compare

    Approve a version and it locks. Analysts, managers, and the CFO compare it against actuals in the same workbook.

Connect the warehouse. Set the drivers. Freeze and compare

Actuals and scenarios

3 scenarios off 1 set of actuals

Change a driver and all 12 periods recalculate, so every scenario stays anchored to the same closed months.

Built for the forecast cycle

Explore

Run a scenario without a handoff

Drivers and overrides live beside the data. Change a growth rate and every downstream period recalculates live, so the analyst who owns the number runs the scenario in one sitting.

  • Adjust a driver and see every period recalculate
  • Compare 2 scenarios side by side on live data
  • Keep overrides and comments inside the app with the data
Act

Freeze a version and prove it

Locking a version freezes it, so nobody overwrites an approved forecast. Every edit carries a user and timestamp, so when two numbers disagree the trail shows who submitted which.

  • Lock an approved version so nobody overwrites it
  • Trace every edit to a user and timestamp
  • Compare an approved version against actuals and reforecast
Govern

Reforecast on a cadence

Automated Actions rerun the forecast on a schedule, and Sigma Agents point at the driver behind a variance. If an agent gets it wrong, the reasoning and the rows are there to check.

  • Rerun the forecast nightly without touching the workbook
  • Flag a variance and name the driver behind it
  • Route AI through the LLM provider you already run
Connections

Always on the newest trusted data

Amazon Redshift
ClickHouse
Databricks
Snowflake

Sigma queries your warehouse live, so a workbook reads what landed a moment ago.

Where the forecast inputs land

Every number a planner types lands in the warehouse under the roles and row-level security already in force.

Change historyAppend only
DWFP&A lead updated Forecast on Jan 202614:22
MEAnalyst edited Assumption on Feb 202614:09
ROSales director approved Scenario on Mar 202613:51
JTFP&A lead reassigned Owner on Jan 202611:04
Append-only writes

Input Tables write to a designated write schema through an append-only write ahead log. Production tables stay untouched.

Inherited permissions

Every query, action, and AI call carries the warehouse roles and row-level security. Each person sees only their own rows.

Full change history

Audit history on the Input Table logs every edit by user, timestamp, and change type. Audit gets its answer without a ticket.

Approved versions stay hardened, down to who changed what

Treasury lead, manufacturing
Every kind of app

Sigma for every kind of app

All 7 build on the same live, governed platform.

Sigma for dashboards

Dashboards that get to the answer

Explore

Sigma for modeling

Model every lever on live data

Explore

Sigma for planning

Own your plan in one place

Explore

Sigma for project tracking

Project tracking you can act on

Explore

Sigma for reconciliation

Reconcile every number in one place

Explore

Sigma for reference data

Fix reference data without a ticket

Explore

Questions finance teams ask

Run the next cycle in Sigma

Bring one forecast cycle into Sigma and watch the versions, approvals, and audit trail hold up under review.