1. Candidate selection
The first 20 names form an initial research pool across UX research, product design, SaaS, enterprise, and specialist work. Profiles remain candidates while source links, operating status, and relevant project-specific evidence are checked.
2. Evidence standards
We prioritize dedicated agency case studies and portfolio pages for the agency’s account of its role; client sources for independent confirmation; review platforms for attributed client feedback; and recognized award bodies for specific awards. Each record preserves the source title, direct URL, publisher, source type, date checked, supported claim, and limitations.
Project relevance
Does a direct source support the work, client relationship, agency role, and category fit?
Attribution
Is the statement agency-reported, client-reported, review-platform data, or independent confirmation?
Freshness and conflict
When was the source checked? Do other sources disagree, and can the conflict be resolved?
3. Scoring and ranking
A proposed 100-point model is under consideration: relevant project evidence (25), category fit (25), documented outcomes (20), independent review evidence (15), source freshness and transparency (10), and buyer-market fit (5). The model is not active: we have not applied it consistently to the full pool or collected project-level score evidence.
Category pages use a disclosed editorial sequence based on the project fit described in each candidate file and the evidence currently on file. This sequence is a navigation aid, not a numerical score, verified performance ranking, or guarantee. UXComparation does not sell placement. No numerical category scores or objective “best agency” claims are published.
4. Category inclusion
Category pages show named candidates in editorial order and link to first-pass source leads and open questions. Their positions are not validated performance rankings. AI and chatbot categories require specific project evidence; deployment status is stated only when a linked source supports it.
5. Reviews and commercial independence
Review ratings are never used alone to determine quality. Any rating must link to its platform and explain the review evidence and its limits. Any future commercial relationship must be disclosed and kept separate from editorial decisions.
6. Updates and corrections
Agency information changes. We record when a source is checked, review material conflicts, and update or remove claims when evidence changes. Readers can report a correction with a direct source at our corrections page.