Phase 3 Search
The Biotech Capital Report · Methodology

The Biotech Capital Report — Methodology & Verification

Phase 3 Search · June 2026

This document explains what the Biotech Capital Report measures, how the underlying data is verified, and what it is — and is not. It is the companion to the published report and exists so readers can understand the standard behind every claim.


What the report measures

A single metric: follow-on rate — the percentage of an investor's biotech portfolio companies in which that investor participated in two or more funding rounds at distinct dates, computed across an 11-year observation window (2015–2026).

Follow-on rate measures conviction, not outcome. A firm that backs a company through Series A, B, and C is making a different bet than one that writes a single check at an IPO-adjacent round. The metric is designed to make that distinction visible.

Tier boundaries: Tier 1 ≥ 50% follow-on · Tier 2 30–49% · Tier 3 < 30%.

Minimum portfolio for inclusion: 5 distinct biotech companies in the window. Firms below the threshold are excluded from the published list.


The dataset

The report is built on the Phase 3 Capital Intelligence dataset — a compiled record of 37,308 biotech venture deal events spanning 1995–2026, covering 2,156 unique companies. The published tier list comprises the 517 investors that meet the five-company minimum within the 2015–2026 analytical window. Earlier rounds are retained in the source but excluded from the follow-on calculation, which requires a multi-year window to register multi-round patterns.

Investor records are reconciled to a canonical entity layer — name variants and casing duplicates of the same firm are merged, while a venture arm and its parent company are kept as distinct actors (e.g., Pfizer Ventures and Pfizer are treated separately, per industry convention).


How the data is verified

The dataset is verified through a multi-pass, multi-source research process that triangulates each fact against primary sources — SEC EDGAR filings (Form D and S-1 principal-stockholder tables), company funding press releases, and investors' own portfolio disclosures. A claim is treated as defensible only when it survives independent verification; records that cannot be verified are tagged and excluded from the published list rather than guessed at.

Before publication, the data passes three independent verification gates:


Sample-size confidence bands

Every firm in the published list carries a confidence band based on portfolio size. Named claims in the body of the report draw only from the Robust and Standard bands; the full interactive table surfaces every firm with its band visible so readers can apply their own threshold.

Band Sample size Count
Robust n ≥ 25 87
Standard n = 15–24 70
Adequate n = 10–14 88
Limited n = 5–9 272

What the data is NOT

Not predictive. Follow-on rate measures historical behaviour across the 11-year window (2015–2026). It does not forecast future investor behaviour.

Not complete. A share of biotech companies have insufficient public disclosure to verify investor lists; those rounds are excluded. The dataset captures the publicly-disclosed universe, not the full venture history.

Not a quality judgment. A firm in Tier 3 is not "worse" than one in Tier 1. Hedge funds, crossovers, and pharma BD entities show low follow-on rates by design — they write single checks at specific moments. Tier 1 and Tier 3 produce similar pooled exit rates; they reach those outcomes via different strategies.

Not investment advice. The report is a structural analysis of capital behaviour. Investment decisions involve many factors not captured by a single metric.


Known limitations and biases

Quiet-participation bias. Firms that participate in rounds as existing investors without being individually named in press releases are systematically under-counted. The published rates reflect the publicly-disclosed pattern. For several prominent firms where this matters, the published rate was cross-checked against the firm's own disclosed portfolio and found consistent.

Crossover under-count. Crossover funds disclose private investments inconsistently; their published low follow-on rates reflect their disclosed pattern, which may understate true private-side activity.

Pharma BD entity boundary. Pharma corporate venture arms are classified separately from the parent pharma company. Direct deals from the parent are tagged Pharma BD; corporate-venture deals are tagged Corporate VC, matching industry convention.


How to audit any claim

The interactive Follow-On Explorer shows every firm with its follow-on rate, sample size, and confidence band. If you believe a specific firm has been mischaracterized, contact us and we will review and correct promptly. The full deal-level source records and verification detail are maintained by Phase 3 Search and available to verified subscribers under confidentiality.


Compiled by Phase 3 Search · June 2026. This document is the methodology companion to The Biotech Capital Report.