API Data Quality Report
It is not enough to say that NinjaPear has good data. Instead, I'd rather be open and show this data sheet for which it will be updated everyday, to tell you:
- the fill rates of each API request - that means given a person/company profile data, how much data do we actually return?
- how fast or slow each API request takes to respond.
- how often are we able to match a person/company to a request.
It is not perfect, but trust me that by being publicly accountable, I'm going to do my best to make sure it improves over time!
- Sample
- 1,000 most recent live lookups per endpoint
- Updated
- Daily
- Last updated
At a glance
Headline numbers per endpoint. Response time, failure rate and match rate are measured on live lookups only.
| Endpoint | Median Response Time | P90 Response Time | 5xx rate | Match rate | Avg. fill rate |
|---|---|---|---|---|---|
| Person profile fast mode | 19.9s | 52.6s | 0.70% | 95.2% | 56% |
| Person profile standard mode | 18.2s | 35.3s | 0.00% | 87.5% | 56% |
| Company details | 17.9s | 32.4s | 9.50% | 95.4% | 75% |
| Company funding | 13.7s | 28.3s | 0.00% | n/a | 58% |
Response time
Time to a complete answer for live lookups that ended in 200 or 404. The large number is p50. The bar spans p10 to p90, labelled at each end, and the tick marks p50.
Server Failure Rate
Share of the most recent live lookups that ended in a server error (5xx). Each bar is one daily report, and all four charts use the same scale. Hover a bar for its rate.
Match rate
Share of live lookups (not served from cache) that found a profile: 200 divided by 200 plus 404. Each bar is one daily report on a 0 to 100% scale.
Attribute fill rate
Share of returned profiles where each attribute is present and not empty, over the last billed profiles per endpoint. Indented rows are measured per list entry, for example per work history role.