Reference Data
Each reference dataset is a full parameter sweep of one of our calculators
across a two-axis grid. Every number you see is the real output of the
shipped engine bundle at /engines/{slug}.js —
recomputed independently in CI and diffed against the published table, so
prose and engine can never drift. Original primary-source data, free to
cite and download as CSV.
Provenance
Computed live from a deterministic engine. The exact input grid is published alongside every table.
Verification
CI re-runs compute() per cell and fails the build on any mismatch. No fabricated numbers.
Reuse
Each dataset ships schema.org Dataset metadata and a one-click CSV export.
Datasets
- Dataset AI Product Margin by Price × Token Cost
Per-user gross margin for a one-person AI SaaS across every combination of monthly price and model output-token cost. Read off the margin floor before you pick a model.
- engine
- AI Product Margin Calculator
- grid
- 8 subscription prices × 8 output-token prices
- cells
- 64
- headline
- Gross margin %
- Dataset LTV:CAC Ratio by ARPU × Customer Lifetime
Margin-adjusted LTV:CAC ratio for a subscription product across monthly ARPU and average customer lifetime, at a fixed $250 CAC and 80% margin. Find the 3× healthy floor before you scale acquisition.
- engine
- Customer Lifetime Value Calculator
- grid
- 7 ARPU tiers × 8 customer lifespans
- cells
- 56
- headline
- LTV:CAC ratio
- Dataset CAC Payback Months by ARPU × Acquisition Cost
CAC payback period in months across monthly ARPU and acquisition cost at 80% gross margin. See exactly where payback crosses the 12-month healthy line before funding growth from cash flow.
- engine
- CAC Payback Calculator
- grid
- 7 ARPU tiers × 8 acquisition-cost levels
- cells
- 56
- headline
- Payback months
- Dataset AI Support Savings by Ticket Volume × Deflection
Monthly cost savings of an AI-first support desk versus humans across ticket volume and AI deflection rate. See where weak deflection actually costs more than staffing agents.
- engine
- AI vs Human Support Cost
- grid
- 6 ticket volumes × 7 deflection rates
- cells
- 42
- headline
- Monthly savings $
- Dataset Embeddings DB Cost by Vectors × Query Volume
Cheapest managed vector-database monthly cost across stored vectors and daily queries, over Pinecone, pgvector, LanceDB and Turbopuffer. The winning vendor flips as you scale.
- engine
- Embeddings DB Cost
- grid
- 6 vector counts × 7 query volumes
- cells
- 42
- headline
- Cheapest $/mo
- Dataset LLM Switching Payback by Spend × Discount
Months for a cheaper model vendor to repay the migration cost, across current monthly spend and the discount offered. See where lock-in is real and where it pays to switch.
- engine
- LLM Vendor Lock-In Cost
- grid
- 6 spend levels × 7 vendor discounts
- cells
- 42
- headline
- Payback months
Looking for the interactive version? Run your own numbers in the calculators, or read how each engine works in the methodology library.