Metrics

Metrics Guide

For detailed definitions of all metrics used in the Termina platform, see our comprehensive Glossary of Metrics.

---

Metric Availability

To view which data is available for a deal, consult the Deal model, for example at the Deal endpoint. Any Deal model will contain metadata in the data field specifying which data exists. The metric data is populated by Termina based on the raw data, financials, and other files uploaded for any given deal.

Within the data field of the Deal object, the following fields specify the requisite parameters to access the various available metrics:

• financials_exist specifies if the financials exist, where the Financials with the deal_id will return the financials metrics response.
• talent_exists specifies if talent/employee data has been processed for this deal.
• product_metrics contain the list of user_type and metric combinations that have been parsed from the raw data, which are then available for query through the Product module.
• unit_economics_metrics contain the list of user_type whose unit economics are available in the Unit Economics module, which require the user_type to have a revenue metric.
  • Note: at this time each user_type is treated independently, and holistic unit margin calculations are not supported.
  • Each user_type is treated as receiving the entirety of acquisition spend. Unit margin can be modeled from these constituents.
• benchmark_user_types contains the list of user types with benchmarks available for this deal. Benchmarks will exist for every user_type that contains a revenue metric.
• benchmark_categories (deprecated) specifies the list of major categories for which a company was labeled. Use the /available-categories endpoint instead for current category information.
• excel_export_exists specifies whether the data export is available for this deal.

---

Benchmarking Data

The Benchmarking module provides comparative analysis against Termina’s proprietary dataset of 1000+ companies. Data is available at three levels:

Global Metadata

General metadata about the benchmarking modeling provided by Termina, accessible via /data/benchmark/metadata/ endpoints:

Account-Level Benchmarks

Endpoints specific to an account/group with benchmarking models tuned to the group’s data partition:

Deal-Level Benchmarks

Benchmarking data for a specific deal within a group:

---

Understanding Benchmark Models

Scaling Models

Scaling models use Gaussian Process regression to show how metrics evolve as companies grow. Each model contains:

• raw_data: Actual data points from the benchmark dataset. Each point is a dictionary with two metrics (e.g., {"rolling_3m_annualized_revenue": 500000, "cmgr12": 0.05}).
• pred: The predicted trend line showing expected values at each scale point.
• bands: Confidence intervals at different significance levels (alpha). A band with alpha: 0.2 means 80% of companies fall within the low-high range.

Available Scaling Models:

Tradeoff Models

Tradeoff models use normal distribution fitting to show the relationship between two metrics at a given revenue scale. Each model contains:

• scale: The revenue range for this tradeoff analysis.
• raw_data: Data points used in the model, each with two metrics (e.g., {"quarterly_operating_margin": -0.25, "cmgr12": 0.06}).
• discarded_raw_data: Outlier points excluded from the model.
• areas: Elliptical confidence regions. An area with alpha: 0.2 contains 80% of companies.

Available Tradeoff Models:

---

Metric Definitions

For detailed definitions of all metrics including CMGR, LTV/CAC, retention rates, and financial metrics, see the Glossary of Metrics.

Key metric categories covered:

• Cohorts: Logo retention, revenue retention, LTV
• Concentration: Gini coefficient, revenue share metrics
• Financials: Revenue, margins, magic number, rule of 40
• Growth Accounting: CMGR, churn, expansion rates
• Talent: Employee metrics, retention, productivity
• Unit Economics: CAC, LTV/CAC ratios