Methodology

Methodology versioning

ValuFai uses a versioned methodology, currently VALUFAI-DPV-0.2. Every valuation run records the methodology version it was produced under, so results from different versions can be distinguished and compared. Version changes are the mechanism by which coefficients and assumptions are improved over time; a new version is a new contract for how numbers are produced.

Deterministic valuation

All numerical outputs — asset values, portfolio adjustment, talent/capability values, ranges, and confidence — are produced by deterministic, auditable calculations. Given the same inputs, the same valuation run is reproducible. The calculations apply documented multipliers and adjustments; they do not rely on an AI model for any number.

AI role separation

Artificial intelligence is used only to assist with non-numeric work: classifying assets, suggesting interview questions, drafting explanatory narrative, and summarizing evidence. AI cannot alter the numerical results. When an AI provider is unavailable, the application continues to function fully with deterministic fallbacks.

Individual vs portfolio valuation

Each asset (website, application, software, domain, IP, business) is first valued on a standalone basis. The portfolio is then valued as a whole, applying a bounded adjustment to reflect how the assets work together — shared customer bases, cross-asset demand, dependencies, and concentration risks. The portfolio result is the primary figure; standalone figures are shown for transparency.

Operating-capability / talent valuation

ValuFai values the assembled team and operating capability behind a portfolio separately from the assets themselves. This reflects replacement cost, time to rebuild capability, retention arrangements, skill scarcity, and key-person concentration. Talent value is modeled as retention-weighted operating capability, not ownership of people. It is reported separately from asset and business value to avoid double-counting.

Evidence states

Every input carries an evidence state — verified, owner-supplied, public source, model-inferred, derived, unavailable, or conflicting. The mix of evidence states drives both confidence and how wide a result range is. Higher-quality evidence narrows ranges; thin evidence widens them.

Low / base / high ranges

Results are expressed as a low, base, and high estimate. The base is the central estimate. The low and high are derived from the base using evidence-dependent spread; they are not a guarantee that a real transaction would land anywhere in the range.

Confidence

Confidence is reported on a 0–100 scale and reflects the quality and completeness of evidence. Confidence is bounded by the evidence actually supplied: unknown information cannot increase confidence, and missing data widens ranges and lowers confidence rather than manufacturing precision.

Unknown evidence

UNKNOWN is a legitimate system state. When information is not provided or cannot be verified, the engine uses conservative default assumptions, widens the range, and lowers confidence. No answer is better than an invented one.

Portfolio premiums and discounts

Portfolio-level value can be adjusted by a bounded amount, expressed in percentage points (for example, −13.75 means −13.75%). Synergy across assets can add value; revenue concentration, platform dependence, and correlated dependencies can reduce it. The adjustment is capped by the methodology limits and is applied uniformly across the range.

Buyer-specific strategic value vs general indicative value

ValuFai distinguishes general-market indicative value from buyer-specific strategic value. Strategic value reflects what a particular acquirer might pay because of unique synergies, replacement-cost savings, time advantages, customer access, capability access, or strategic fit. It is modeled separately and labeled with its assumptions, and it may differ materially from general-market value — it can be higher for a buyer that realizes those benefits, and it can be lower for a buyer that does not. Strategic value is not part of the general indicative range unless explicitly presented as buyer-specific.

Limits of the methodology

ValuFai estimates indicative value based on the information entered and on general small-digital-business market conditions. It does not and cannot know the price a particular buyer would pay. Multipliers are working assumptions that are intended to be re-fit against real transaction data over time. Results should not be relied upon for a transaction without independent professional advice.

Last updated: 2026-08-08.