Methodology 2026.10
Scores are computed by a deterministic engine from facts stored in git. Nobody types a score, a rank, or a pro/con by hand. The same rules apply to every tool, whoever makes it.
Evidence ladder
- official-docs / official-other: 100% credit
- third-party-confirmed: 100% credit
- vendor-claimed (marketing without documentation): 50% credit
- not-verified: unknown. Unknown is not zero: the criterion scores the catalog average for that criterion (the mean over tools where it is known), and coverage shows how much of the score rests on known facts.
Profiles
| Component | General | By tool type | Team / agency |
|---|---|---|---|
| Safety & Account Health | 28 | 35 | 25 |
| Multichannel & Data Quality | 25 | 20 | 30 |
| Pricing & Value | 18 | 15 | 15 |
| AI & Developer Readiness | 15 | 15 | 15 |
| Reputation | 10 | 10 | 10 |
| Free Tier / Trial | 4 | 5 | 5 |
Leaf weights
Safety & Account Health
dedicated_proxy 25 · limits_safety 30 · account_rotation 15 · runs_offline 20 · pending_cleanup 10
Multichannel & Data Quality
multichannel_native 30 · li_actions 20 · unified_inbox 10 · email_finder_quality 30 · sales_navigator 10
Pricing & Value
pricing_transparency 40 · price_value 60
AI & Developer Readiness
mcp 40 · api 30 · openapi 10 · webhooks 10 · zapier 10
Reputation
reviews 100
Free Tier / Trial
free_access 100
Key rules
dedicated_proxy: dedicated IP per account = 1, shared or bring-your-own = 0.4, none = 0.limits_safety: share of randomized delays, daily caps, SSI monitoring and warm-up that are supported.multichannel_native: email + LinkedIn in one sequence = 1, via integration = 0.5, none = 0.email_finder_quality: finder 40, verification 30, database size 30 (≥100M contacts = full, ≥10M = 2/3, smaller = 1/3).mcp: official remote = 1, official local = 0.7, community = 0.3, none = 0.price_value: price for one account vs the catalog median ($79). Median = 0.5, half the median or less = 1.reviews: mean rating across G2, Capterra and Trustpilot weighted by log(1 + review count); 3.0★ = 0, 5.0★ = 1.free_access: free forever = 1, trial ≥14 days = 0.7, shorter trial = 0.5, none = 0.
A ranking is published only when at least 5 tools are rated.
Changes to this methodology
- 2026-10-15: Initial public methodology. Weights per spec section 5.1.
- 2026-10-15: Unknown facts now score the catalog average for that criterion instead of being dropped with weights renormalized. Dropping them let sparsely documented tools outscore well-documented ones (a tool could score 100 on a component from a single known fact). Applies to every tool; weights unchanged.
Machine-readable: /api/v1/methodology