How scores work
Every product gets one critic score from 0 to 100: a weighted average of professional reviews. It uses the same method as Broadway Scorecard, adapted for tech outlets.
1. Each review becomes a number
A published rating sets a range, the same rule Broadway Scorecard uses for star ratings: 4.5/5 or 9/10 and up lands between 91 and 100, 3.5/5 to 4/5 (7/10 to 8.9/10) between 71 and 90, and lower ratings in lower ranges. Two AI models (from OpenAI and Google) then read the full review and place it inside that range: near the top when the reviewer is enthusiastic, near the bottom when the rating comes with big reservations. The review's score is the average of the two, and it can never leave the range the outlet set. When an outlet publishes no rating, the same two models score the verdict from 0 to 100, and the review is labeled “Unrated, scored from text.”
2. Bigger outlets count for more
- T1100%Major national outlets with full-time reviews staff (The Verge, WSJ, CNET, Engadget, Wired)
- T275%Large dedicated tech publications (Tom's Guide, TechRadar, PCMag, ZDNET, Digital Trends)
- T340%Specialist outlets and general-interest publications (GSMArena, Pocket-lint, Popular Mechanics, Daily Mail)
- T420%Blogs, newsletters and smaller outlets
A product needs at least 5 reviews before it gets a score.
3. The score sets the label
- 85Critical Gold85-100Best in class. If you are buying one, buy this.
- 75Recommended75-84A strong buy for most people.
- 65Worth a Look65-74Good, with caveats. Best if its strengths are your priorities.
- 55Mixed55-64Critics are split.
- 0Critical Miss0-54Not recommended.