Mission
PhaseFolio provides an audit-grade methodology framework for early-stage biotech asset valuation. The research engine combines peer-reviewed clinical success rate data with version-controlled risk-adjusted NPV methodology, producing institutional-quality pipeline valuations that are reproducible, independently verifiable, and defensible to investment committees, limited partners, and boards.
Every assumption is traceable. Every benchmark is sourced. Every calculation is documented in our methodology.
Model Governance
A valuation is only as trustworthy as the process behind it. PhaseFolio governs its methodology the way a research institution governs a model — sourced inputs, documented derivations, independent verifiability, and disciplined version control. The same discipline runs through the platform and its security posture.
Every probability-of-success benchmark derives from peer-reviewed, published sources — never proprietary data or expert-opinion priors. The foundational dataset is the BIO/QLS/Informa Clinical Development Success Rates study (2011–2020); the complete source list — every benchmark and multiplier study, with citations and DOIs — is maintained in the methodology references.
Multiplier adjustments (genetic validation, orphan designation, biomarker strategy, mechanism novelty) are implemented as odds ratios using logistic transformation, preserving the [0,1] probability bound. Sources and derivations are documented in Section 1.2 of the methodology.
The rNPV engine is validated against published academic benchmarks. Phase-transition rates are cross-referenced with BIO/QLS 2021, Wong et al. 2019, and Zhou et al. 2025. Monte Carlo simulations use sequential Bernoulli (rpNPV) sampling, with Gaussian copula correlation for cost-duration uncertainty.
Beyond benchmark cross-referencing, the methodology is validated out-of-sample. A retrospective backtest measures how well the engine’s risk-adjusted scores separate eventual approvals from failures across curated drug cohorts — oncology, rheumatoid arthritis, and antimicrobials — reported as AUC with Wilson confidence intervals.
Backend computation is covered by automated tests that verify output consistency against known analytical solutions and published reference values.
Every PhaseFolio export is cryptographically signed and embeds the exact engine version, methodology version, and benchmark dataset versions used to produce it. An investment committee, an auditor, or an AI diligence agent can confirm an export’s integrity and provenance through the public verify endpoint — no account required.
Signatures remain verifiable indefinitely. A methodology update never invalidates a previously issued export — each one stays reproducible against the methodology in force when it was signed.
Methodology, engine, and benchmark datasets are versioned independently. Benchmark data is updated when new peer-reviewed studies are published; the CT.gov and FDA data pipelines provide supplementary real-world evidence for phase-transition validation. Every update is versioned and documented, so any valuation can be reproduced against the methodology in force when it was issued.
Current methodology version: methodology@2026-06-17. See the full methodology for the complete specification.
Team
Tamal Adebisi
Founder & Methodology Author
MS Finance candidate at Johns Hopkins University. Authored the PhaseFolio rNPV methodology, the multi-indication backtest framework, and the engine architecture and clinical data pipeline that underpin the published valuations.
LinkedIn ↗Jeromie Ballreich, PhD
Director, M.S. in Health Economics & Outcomes Research
Associate Research Professor in the Department of Health Policy and Management at Johns Hopkins Bloomberg School of Public Health. His research spans pharmaceutical economics and the economic evaluation of drugs alongside clinical evidence; he advises PhaseFolio on valuation methodology and benchmarking.
Data & Privacy
PhaseFolio processes financial projections that may be commercially sensitive. All project data is scoped by organization and encrypted in transit (TLS 1.3); every request is filtered to the authenticated organization so one organization's data is not exposed to another. Authentication is handled by a managed identity provider using organization-scoped access tokens.
We do not sell, share, or use customer data to train models. Benchmark data is derived exclusively from publicly available, peer-reviewed clinical trial databases and regulatory filings.
Contact
For methodology questions, partnership inquiries, or feedback, reach us at contact@phasefolio.com.
PhaseFolio is currently in its Design-Partner Phase. Feedback from early users is actively incorporated into methodology and engine development.