A16z's metrics essay defines terms, it does not benchmark them

The record

Andreessen Horowitz partners Jeff Jordan, Anu Hariharan, Frank Chen and Preethi Kasireddy published '16 Startup Metrics' in August 2015, and Hariharan, Chen and Jordan followed with '16 More Startup Metrics' the following month. As retrieved on 16 September 2026, both essays remain posted on a16z's site as reference material rather than a dated news item. The first essay defines terms including bookings, as 'the value of a contract between the company and the customer'; monthly and annual recurring revenue; customer acquisition cost, 'the full cost of acquiring users, stated on a per user basis'; gross merchandise value; and churn, defined as lost customers divided by the prior month's total.

What the documents establish

The essay itself states its purpose is definitional: it exists so operators use the same vocabulary when discussing their own numbers, not to grade a company against peers. Neither essay presents original survey data or a comparable-company dataset; the first is a glossary with worked definitions, and the second, while still built around definitions such as total addressable market and net promoter score, does cite specific figures for named public companies, including a stated Facebook average revenue per user for a reported quarter, to illustrate how a metric is calculated in practice. Those illustrative figures come from the companies' own disclosures at the time, not from a16z's portfolio or any survey a16z conducted, and the essays should not be cited as if they were a16z's own benchmark study.

The operating read

The practical value of the essays is standardizing vocabulary before a number gets compared to anything: a 'churn rate' quoted without saying whether it is customer count, logo, or revenue churn, or monthly versus annual, cannot be benchmarked against another company's figure even if both call it churn. Editorially, a founder preparing a fundraising deck can use the essay's definitions as a checklist for internal consistency, confirming that MRR excludes one-time fees and that CAC includes the fully loaded cost of the team doing the acquiring, before ever comparing the resulting number to an outside benchmark report.

What to check before you decide

Before presenting a metric using these definitions, check the following.

• Does your calculation match the essay's definition exactly, including which costs and revenue types are included or excluded?
• Are you treating an illustrative company example from the second essay as if it were a16z's own benchmark data?
• Has the metric's definition drifted internally since your last board deck, even if the label on the slide has not changed?

These essays are a shared vocabulary, not a scorecard, and their usefulness depends entirely on applying the definitions consistently before any comparison is attempted.