How we rank

The Consensus Score (100 points)

45

Editorial consensus. We count the distinct trusted outlets that have recommended the product. Tier 1 outlets contribute 1.0, tier 2 contribute 0.6, tier 3 contribute 0.3. The weighted sum is capped at 3.0 and scaled to 45 points. Two Wirecutter and Strategist picks will beat a long tail of lower-tier roundups.

20

Quality signal. When a credible publisher reports an aggregate rating and review count, we blend it with a conservative prior (50 ratings at 3.9 stars) and normalize the resulting Bayesian average to this bucket. If no credible rating exists, we award a neutral 10.

25

Persona fit. A separate model rates each candidate 0 to 10 against the persona's description and price band. We scale that to 25.

10

Longevity. Mentioned by trusted outlets in two or more distinct years: 10. Only mentioned in the current year: 6. Otherwise: 3.

-

Penalties. Apparel -10, consumable -8, size-dependent -10, price outside the persona's band -15. Gift cards, non-products, and anything without a trusted source are excluded.

Scores are clamped to the 0 to 100 range. The complete per-pick breakdown is stored and surfaced on every list.

Diversity rule for Top 5

After scoring, we greedily pick the Top 5 subject to three guardrails: at most one product per brand, at most two per subcategory, and - once candidates allow - at least two different price tiers represented by pick four and three tiers by pick five. The goal is a Top 5 that looks like a shortlist, not five variants of the same gadget.

Trusted sources

Only these domains contribute to the Consensus Score. Everything else is weight zero.

What we do not do

• We do not read Amazon customer reviews or use any Amazon API.
• We do not accept paid placements or sponsored picks.
• We do not display product images, prices, or star ratings on list pages.
• We do not collect user data or accept submissions.