Firecrawl Company Enrichment API Benchmark 2026 Results
How Data Legion company enrichment scores on Firecrawl's Company Enrichment API Benchmark 2026, beside every provider it ranks.
Our run Nov 16, 2026 · 200 companies · 6 providers
How does Data Legion score on Firecrawl's benchmark?
On Firecrawl's Company Enrichment API Benchmark 2026, Apollo.io leads the common fields (89.3%) and Data Legion scores 80.0%, 3rd of 6, with 97.7% on the LinkedIn page and 85.8% on headquarters.
Firecrawl Company Enrichment API Benchmark 2026
Six providers enrich the same 200 US companies, and code scores headquarters, employee count, founding year and LinkedIn page (the common fields), plus industry, ticker and revenue, against SEC filings, federal records and company websites, frozen before any provider was called.
- Apollo.io89.3
- People Data Labs80.6
- Data Legion80.0
- Ocean.io78.7
- FullEnrich72.7
- ContactOut72.7
Headquarters, employee count, founding year and LinkedIn page, scored together.
- Apollo.io81.6
- People Data Labs72.0
- Data Legion70.6
- Ocean.io64.9
- ContactOut60.6
- FullEnrich59.6
The common fields plus industry, stock ticker and revenue.
- Apollo.io91.0
- People Data Labs88.7
- Data Legion85.8
- Ocean.io85.0
- FullEnrich75.7
- ContactOut64.7
Same city as the filings, or another city in the same metro area.
- Apollo.io77.2
- Ocean.io56.5
- Data Legion56.0
- ContactOut54.6
- FullEnrich54.1
- People Data Labs51.8
Full credit within 1.2 times the filed figure, half credit within 2 times.
- Apollo.io88.7
- People Data Labs85.0
- Data Legion83.8
- Ocean.io80.8
- ContactOut78.9
- FullEnrich67.6
Within one year of the founding year on record.
- Apollo.io99.2
- People Data Labs98.8
- Data Legion97.7
- ContactOut95.9
- Ocean.io94.4
- FullEnrich93.2
The company's own LinkedIn page.
- Ocean.io69.2
- ContactOut65.2
- FullEnrich62.7
- People Data Labs62.5
- Data Legion61.1
- Apollo.io53.4
Industry label graded 1 for a best fit, 0.75 for a broader or secondary label, 0 otherwise.
- Apollo.io91.3
- People Data Labs90.0
- Data Legion88.2
Share of public companies where the provider returns one of the company's tickers. A dash: not offered.
- Apollo.io82.8
- ContactOut7.6
- People Data Labs4.8
- Ocean.io2.1
Share of public companies where the provider's revenue is within 5% of the latest fiscal year. A dash: not offered.
| # | System | Common FieldsHigher Is Better | All FieldsHigher Is Better | HQHigher Is Better | EmployeesHigher Is Better | FoundedHigher Is Better | LinkedInHigher Is Better | IndustryHigher Is Better | TickerHigher Is Better | RevenueHigher Is Better |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Apollo.io | 89.3 | 81.6 | 91.0 | 77.2 | 88.7 | 99.2 | 53.4 | 91.3 | 82.8 |
| 2 | People Data Labs | 80.6 | 72.0 | 88.7 | 51.8 | 85.0 | 98.8 | 62.5 | 90.0 | 4.8 |
| 3 | Data Legion1, 2 | 80.0 | 70.6 | 85.8 | 56.0 | 83.8 | 97.7 | 61.1 | 88.2 | — |
| 4 | Ocean.io | 78.7 | 64.9 | 85.0 | 56.5 | 80.8 | 94.4 | 69.2 | — | 2.1 |
| 5 | FullEnrich | 72.7 | 59.6 | 75.7 | 54.1 | 67.6 | 93.2 | 62.7 | — | — |
| 6 | ContactOut | 72.7 | 60.6 | 64.7 | 54.6 | 78.9 | 95.9 | 65.2 | — | 7.6 |
- 1Our row is our own run of the benchmark's scoring code (v1.0.1) and scorer on its published 102-company sample, against Data Legion company enrich on production. The published leaderboard uses all 200 companies, so the rows are close but not identical samples.
- 2The benchmark's scoring code does not read Data Legion's headquarters field (it sets it to empty), so Firecrawl's published leaderboard scores our headquarters as zero and shows us at 52.2% common, 54.1% all. Our row reads the headquarters from the primary office in `locations`, where company records carry it, and changes nothing else.
Other systems' scores are as the benchmark published them on Oct 2, 2026. Built and published by Firecrawl, which resells four of the six providers, Data Legion among them, and says so on its page. Benchmark page, opens in a new tab · Code and data, opens in a new tab
How It Is Measured
The same input for every system, scored by the benchmark's own rules.
The same companies for every provider
Every provider gets the same 200 companies by domain, public and private, and the reference values were frozen before any provider was called.
Scored by code
Answers are compared with SEC filings, federal records and company websites by code rather than a model, so the same answers always get the same score.
Confidence intervals on every score
Every score carries a 95% confidence interval, and each field reports its coverage and its accuracy when answered.
Our run on the public sample
Firecrawl publishes 102 of its companies with its scorer. Our row is our run on that sample with Firecrawl's own scorer, as notes 1 and 2 describe.
Common Questions
What buyers ask when they compare providers.
Firecrawl's scoring code sets Data Legion's headquarters field to empty, so its published leaderboard counts every one of our headquarters answers as wrong. Our run reads the headquarters from the primary office in each company record and changes nothing else in the scoring: 80.0% on the common fields and 85.8% on headquarters. It also uses Firecrawl's published 102-company sample and our production API, where the leaderboard used all 200 companies through Firecrawl's own route, so our other fields differ a little from it too.
On Firecrawl's Company Enrichment API Benchmark 2026, Apollo.io leads the common fields (89.3%) and Data Legion scores 80.0%, 3rd of 6. Common fields: Apollo.io 89.3%, People Data Labs 80.6%, Data Legion 80.0%, Ocean.io 78.7%, FullEnrich 72.7%, ContactOut 72.7%.
Enrich the same list of your own companies with each provider, then check the fields you will use against a source you trust (filings, the company's own site): headquarters, employee count, founding year, industry and LinkedIn page. Count a blank answer as a miss.
Yes. Firecrawl publishes its code, scorer and a 102-company sample on GitHub, linked above, and any provider can be run against them.
Test it on your own data
Run your own searches and lookups before you buy. A lookup that finds no match costs nothing.