Estimator overview
Estimator lets you build parameterized probabilistic models — anything from a consulting cost estimate to a hiring plan, TAM/SAM/SOM sizing, product NPV, or marketing spend break-even analysis. You write a model once and reuse it across many concrete cases, each stored with its input values and a snapshot of results.
What it does
You define a model using either:
- Formula engine — a mini-language expression where you declare random variables and derived calculations, then run 1,000 simulation iterations to see the output distribution.
- Spreadsheet engine (Pro) — upload an
.xlsxworkbook, mark assumption and output cells, and run the same simulation against your real spreadsheet.
A sketch (no user inputs) works like a one-off calculator — write a model and read off the results. A template (one or more user input fields) is reusable: a runner fills in the fields and saves each case as a named run. All runs for a template are stored together with the inputs and output metrics for comparison and export.
Who this is for
- Analysts and consultants building repeatable cost or revenue models they hand off to clients or colleagues to fill in.
- Finance and ops teams who want to capture uncertainty in a forecast rather than pretending a single number is right.
- Product managers sizing TAM, pricing scenarios, or feature adoption ranges.
- Engineers estimating latency budgets, failure rates, or capacity numbers with known uncertainty.
Layout
The tool has four screens:
| Screen | Path | Purpose |
|---|---|---|
| Author view | /app/estimator/<id>/ | Build and edit the model (formula textarea + derived cards, or spreadsheet grid) |
| Template detail | /app/estimator/<id>/runs | View and compare all saved runs for a template |
| Run view | /app/estimator/<id>/runs/<run_id>/ | Fill in a run's inputs and see results |
| Draft run | /app/estimator/<id>/runs/new | Create a new unsaved run against a template |
For templates, the author view and run view are linked by a Template / Runner toggle in the toolbar.
Next steps
- Core concepts — sketches, templates, runs, engines, and the variable model
- How-to guides — task-oriented walkthroughs
- Reference — mini-language syntax, distributions, metric types, and API endpoints