Research methodology

# How this directory researches and scores software

The score is a shortlist aid, not a verdict. Every published vendor is assessed against the same capability model, and the page keeps missing public evidence visible instead of turning it into a guessed feature or price.

## Sources and verification

Research starts with current vendor product, pricing, security, integration, documentation, and company pages. Trusted third-party review sources can add user-review context. Each volatile fact carries the date it was checked when that date is available.

Vendor submissions do not guarantee publication, placement, or a score change. No vendor can buy a favorable conclusion. Corrections should include the page, the disputed statement, and a current source.

## The capability model

Each vendor receives a 0 to 3 editorial assessment across nine jobs: RFP response, DDQs, security questionnaires, AI drafting, answer-library depth, source attribution, workflow and approvals, integrations, and self-serve onboarding.

- **0:** not a documented focus in the reviewed evidence.
- **1:** limited or adjacent support.
- **2:** capable support for the job.
- **3:** a core, well-documented part of the product.

The visible 5-point score maps the sum of those nine assessments onto a 3.0 to 5.0 review-site scale. It is deterministic: the same capability values that power the comparison tables produce the score. It is an editorial assessment, not a customer rating or certification.

## How category and workflow pages are assembled

A vendor appears in a category when its published record names that category. Workflow groups use a stated capability threshold, then link back to the profile where the reader can inspect pricing, evidence, limits, and the last verification date. Default lists run alphabetically; an interactive sort does not change the underlying evidence.

## How alternatives and comparisons are selected

Alternatives are ordered by job overlap: shared categories, shared tags, aligned capabilities, and pricing model. Comparisons are published only for live pairs in the directory. Neither method declares one universal winner. The useful answer is which product fits a stated workflow, team, evidence standard, and buying model.

## Freshness and unknowns

Pricing, certifications, integrations, product packaging, and company details change. Pages show a check date when the vendor record carries one. If no reliable date is present, the page tells the reader to confirm the current detail with the vendor.

“Not publicly documented” means the research did not find enough current public evidence. It does not mean the product lacks the feature. This distinction matters when comparing private pricing, security controls, APIs, or deployment options.

## Corrections and contact

Send corrections to [corrections@rfp-software-tools.com](mailto:corrections@rfp-software-tools.com). Include a current first-party source where possible. The editorial team reviews material changes before updating the vendor record, page copy, and check date.