{{leaderboard?'Enrichment Leaderboard':'Enrich Arena'}}
People Search Bench
How well do agents find the right people?
Published results from treg’s people-search benchmark ↗
{{benchSelected.name}}
Published score · higher is better
Scores compare complete search systems across a use case. They do not measure an individual enrichment vendor’s hit rate or email-verification accuracy.
Across every use case
The same scale, side by side.
About these results and sources
These charts reproduce the category scores published on treg’s people-search landing page. They are a published snapshot, not a new benchmark run. No overall score is calculated here.
The original People Search Bench evaluates relevance, coverage and information utility. Its current GitHub results differ from the figures on the landing page; use the linked methodology for those definitions and results. A score is not a guarantee that a returned email can receive mail.
{{taskLabel(taskId)}}
Which vendors return the data you need?
{{chartViews.find(v=>v.id===statsView)?.label}} {{variantLabel(inputKeys)}}
{{statsView==='verified'?(taskId==='people.email.find'?'Valid among sampled emails · higher rate to the left':'Reachability and ownership checks unavailable'):statsView==='price'?'Estimated price per entry · cheapest first':statsView==='speed'?'Median response time · lower is faster':statsView==='value'?'Lower price ← · ↑ higher '+(insightVerification?'verdict rate':'hit rate'):(insightVerification?'Returned verdicts':'Returned data')+' · higher rate to the left'}}
{{verifiedExplanation}}
Bar length shows relative price · USD per entry
No vendor prices available for this selection yet.
{{statsView==='verified'?(taskId==='people.phone.find'?'Phone reachability and ownership have not been verified.':'No comparable verification samples for this selection yet.'):statsView==='price'?'No vendor prices available for this selection yet.':insightsState==='loading'?'Loading historical stats…':insights?.status==='warming'?'Processing stored call history…':'No comparable samples for this selection yet.'}} View vendor prices in Arena →
No vendors have both a rate and price for this selection yet.
{{verifiedRateNote(chartDetail)}}
Current catalog estimates per entry. Billing may be per request or per successful result; final pricing depends on your team and query.
Different vendors received different inputs. These are observed results, not a controlled benchmark{{insightVerification?' or verification accuracy':''}}.
| Vendor | {{insightVerification?'Verdict rate':'Hit rate'}} | Response time | Price · {{inputRows.length}} entries | |
|---|---|---|---|---|
{{insightRate(p)}} {{insightSampleLabel(p)}} |
{{verifiedRate(p)}} |
{{insightResponseTime(p)}} median · shorter is faster |
{{p.estimate_micro==null?'Price unavailable':usd(p.estimate_micro)}}{{billingLabel(p)}} | |
| {{customServices?'Click a vendor avatar to include it.':'No services support this input type.'}} |
{{readyQuote?(!leaderboard&&mode==='waterfall'?'Quoted order · stops at the first hit.':'Estimates for your team.'):'Catalog estimates · access and final pricing depend on your team and query.'}}
Loading historical stats…
Stats refresh unavailable; any displayed values are from the last update.
About these stats {{insightWindow}}
Verdict rate is the share of unique evaluable requests that returned a verdict, including valid, invalid and unknown answers. It does not measure verification accuracy.
Hit rate is the share of unique evaluable lookups that returned the task’s required fields. We keep the latest evaluable response per identical request for each vendor, so repeat calls do not dominate. Under 20 unique lookups shows no rate; under 100 is a small sample.
Known bad requests, rate limits, balance/credit failures, access and service failures, refusals and pending responses are excluded. Genuine no-result responses count as misses; ambiguous or missing evidence is omitted. Response time is the median of successful calls, including repeats, with at least 20 returned results.
{{verifiedExplanation}} Verifier details and check dates are available on hover. Samples with fewer than 20 completed checks are withheld. Risky, unknown and conflicting judgments are not counted as valid; unfinished checks are excluded.
Processing stored call history. Stats will appear when the first collection finishes.
These stats use calls with treg’s credentials over a rolling 30-day window. Updated {{insightReviewedOn}}. Vendors received different inputs and waterfall positions, so these figures are not a controlled ranking or verified accuracy. Prices are current estimates for your selection.
Download aggregate data ↗This run did not finish. Completed answers are kept; uncertain calls will not be retried automatically.
| Vendor | {{discoveryRun()?'Queries with matches':run.mode==='waterfall'?'Additional results':'Found'}} | Feedback | {{discoveryRun()?'Cost / kept query':'Cost / kept'}} | Median time |
|---|---|---|---|---|
| {{run.mode==='compare'?Math.round(v.found/v.total*100)+'% · ':''}}{{v.found}} / {{run.mode==='compare'?v.total:v.attempted}} attempted | {{v.up}} 👍 · {{v.down}} 👎 | {{usd(v.costPerKept)}}{{usd(v.cost)}} total | {{v.median==null?'—':seconds(v.median)}} | |
{{run.mode==='waterfall'?'Later vendors receive only unresolved entries; their hit rates are not directly comparable.':'Coverage crowns use results kept after thumbs-down rejections.'}} {{discoveryRun()?'Coverage counts queries with matches, not the number or relevance of contacts.':'Found is not verified accuracy.'}} Timing excludes queue time.
- {{running||runCostSummary.pending?'Total cost so far':'Total cost'}}
- {{usd(runCostSummary.total)}} · Finalizing charges
- {{discoveryRun()?'Average per matched query':'Average per result'}}
- {{usd(runCostSummary.average)}} · {{runCostSummary.found}} {{discoveryRun()?'queries matched':'entries found'}}
| Entry | |
|---|---|
|
|
|
| No entries match this filter. |
{{run.stop_reason}}