How HypeMetrix Calculates Scores
HypeMetrix aggregates opinions from multiple sources, analyzes sentiment with AI, and blends everything into a single, transparent score.
Data Sources
We pull from three distinct source types to capture the full spectrum of user and expert opinion.
Reddit
Real user discussions, complaints, and praise from enthusiast communities.
YouTube
Detailed video reviews from creators and tech channels.
Editorial
Written reviews and articles from tech publications and blogs.
The Scoring Pipeline
For each source, we score individual criteria (sound, comfort, battery, build, etc.) on a 0–100 scale. Each criterion is weighted by its importance for that product type.
We calculate a per-source overall score. If a source misses some criteria, we normalize the result so the product isn't unfairly penalized.
Finally, we blend per-source scores using global source weights and a confidence multiplier. Sources with limited data contribute less; unavailable sources are excluded entirely.
Formula Summary
sourceScore = sum(criterionScore × criterionWeight) / totalWeightFound
blendedScore = sum(sourceScore × sourceWeight × confidence) / sum(sourceWeight × confidence)
- Reddit weight: 45%
- Editorial weight: 35%
- YouTube weight: 20%
Confidence Levels
Data quality affects how much a source contributes to the final score.
High
All major sources had sufficient data for a reliable analysis.
Medium
At least one source had limited data. Results may vary slightly.
Low
Most sources had limited or no data. Treat scores with caution.
Limitations & Transparency
Our scores reflect aggregated public opinion at a point in time. They are not a substitute for hands-on testing, and source availability can vary by product and language.