investorstack
Run the check

METHOD / V0.1

Evidence in. Explicit rules. No runtime model call.

InvestorStack evaluates a bounded operating decision for one firm context. It does not claim a universal software ranking.

Run the private check

001 / DECISION PIPELINE

The abstention is part of the product.

Unknown incumbents, stale provenance, and weak fit can produce HOLD or INSUFFICIENT EVIDENCE. The engine does not fill silence with editorial certainty.

  1. 01Validate the input

    Firm shape, ranked workflows, incumbent health, ownership, budget, security, integrations, and replacement appetite.

  2. 02Resolve eligibility

    Remove candidates that conflict with operating constraints before any score can influence a recommendation.

  3. 03Weight the fit

    Use seven evidence dimensions to order the remaining candidates for this context only.

  4. 04Attach the boundary

    Every product recommendation carries confidence, reviewed sources, assumptions, and unresolved questions.

002 / DIMENSIONS

Internal weights, visible evidence.

Human Usability
20
Workflow Fit
20
Ai Readiness
20
Integration Surface
15
Data Portability
10
Roi Potential
10
Vendor Maturity
5

Dimension values are inputs to a context-specific rule set. InvestorStack does not show their weighted sum as an objective category score.

003 / CONFIDENCE

Freshness changes the claim.

High confidence requires at least two independent tier-one or tier-two sources reviewed within 90 days. Missing provenance fails closed. Tool facts and price bands remain directional until verified directly in a buying process.

Research and scoring run offline against versioned sidecars. The production request path reads committed artefacts and produces the same content for the same input, dataset, method, reference, and timestamp.

Commercial boundary

Paid placement, introductions, and advisory work cannot create, remove, or reorder evidence. Corrections enter an editorial review rather than changing a live score automatically.