Company Enrichment API
Benchmark 2026
Compare company enrichment APIs on how accurately they fill in a company's profile across seven fields, from nothing but its website.
- Last updated
- [ Updated Oct 02, 2026 ]
- Systems compared
- [ 6 providers ]
- Dataset
- [ 200 companies ]
How the providers stack up
Which company enrichment API is best?
Apollo.io is the most accurate company enrichment API in this benchmark. It scores 89.3% on the four common fields, 8.7 points ahead of People Data Labs, and has the highest score on 6 of 7 fields. The exception is industry, where Ocean.io leads.
| Use case | Best pick | Measured result |
|---|---|---|
| Filling in CRM company records |  Apollo.io | 89.3% on the four common fields (HQ, employees, founded, LinkedIn), 8.7 points ahead of People Data Labs |
| The most complete company profile | Apollo.io | 81.6% across all seven fields, 9.6 points ahead of People Data Labs |
| Employee counts for sizing and routing | Apollo.io | 77.2%, 20.7 points ahead of Ocean.io, the next best |
| Revenue for public companies | Apollo.io | 82.8% within 5% of the filed figure; no other provider scores above 8% |
| Industry labels for segmentation |  Ocean.io | 69.2%, 4.0 points ahead of ContactOut; Apollo.io ranks last at 53.4% |
| Finding a company's LinkedIn page | Apollo.io People Data Labs | 99.2% and 98.8%, effectively tied |
| A second source alongside Apollo.io | People Data LabsOcean.io | 80.6% and 78.7% on the common fields, statistically tied for second |
Disclosure: Firecrawl runs this benchmark and sells access to four of the six providers tested (Apollo.io, People Data Labs, FullEnrich and Data Legion) through Firecrawl Alexandria. A public sample of the answer key, the harness and the scoring are published on GitHub.
Leaderboard
Overall score with a 95% confidence interval, on the four common fields or all seven. Switch fields to see how the ranking changes.
Check all metric definitionsApollo.io[alexandria]People Data Labs[alexandria]Ocean.io
FullEnrich[alexandria]
ContactOut
Data Legion[alexandria]Apollo.io[alexandria]People Data Labs[alexandria]Ocean.ioFullEnrich[alexandria]ContactOutApollo.io[alexandria]Ocean.ioContactOutFullEnrich[alexandria]People Data Labs[alexandria]Data Legion[alexandria]Apollo.io[alexandria]People Data Labs[alexandria]Data Legion[alexandria]Ocean.ioContactOutFullEnrich[alexandria]Apollo.io[alexandria]People Data Labs[alexandria]ContactOutData Legion[alexandria]Ocean.ioFullEnrich[alexandria]On the four common fields, Apollo.io leads by 8.7 points. Apollo.io (89.3%) ranks first, People Data Labs (80.6%) and Ocean.io (78.7%) are statistically tied for second, FullEnrich (72.7%) and ContactOut (72.7%) are statistically tied for fourth, and Data Legion (52.2%) is last. Every other gap is statistically significant. Gaps are tested with a paired bootstrap over companies and a Holm correction across all 15 pairs of providers.
Full results
Every field for all 6 providers. The best value in each column is highlighted. Columns marked extra are left out of the headline score, and "not offered" means the provider does not return that field, which counts as 0.
| # | Provider | HQ (higher is better) | Employees (higher is better) | Founded (higher is better) | LinkedIn (higher is better) | Industry (higher is better)extra | Ticker (higher is better)extra | Revenue (higher is better)extra | Common fields (higher is better) | All fields (higher is better) |
|---|---|---|---|---|---|---|---|---|---|---|
| 01 | Apollo.ioalexandria | 91.0% | 77.2% | 88.7% | 99.2% | 53.4% | 91.3% | 82.8% | 89.3% | 81.6% |
| 02 | People Data Labsalexandria | 88.7% | 51.8% | 85.0% | 98.8% | 62.5% | 90.0% | 4.8% | 80.6% | 72.0% |
| 03 | Ocean.io | 85.0% | 56.5% | 80.8% | 94.4% | 69.2% | not offered | 2.1% | 78.7% | 64.9% |
| 04 | FullEnrichalexandria | 75.7% | 54.1% | 67.6% | 93.2% | 62.7% | not offered | not offered | 72.7% | 59.6% |
| 05 | ContactOut | 64.7% | 54.6% | 78.9% | 95.9% | 65.2% | not offered | 7.6% | 72.7% | 60.6% |
| 06 | Data Legionalexandria | not offered | 46.0% | 84.3% | 95.9% | 63.7% | 89.3% | not offered | 52.2% | 54.1% |
How is company enrichment measured?
Same input for every provider
Each provider gets only the company's website, in one lookup per company.
Answer key built first
Every value comes from SEC filings, federal records or the company's own website. The key was frozen before any provider was called; every change made after the calls is listed in the repository.
Seven fields, scored by fixed rules
Code scores six fields, and every industry label on the lists picked before any provider was called. An industry label outside those lists goes to five AI models and the middle grade counts, which is why industry is left out of the headline.
Two leaderboards
Common fields, the headline, covers the four fields scored by code alone: headquarters, employee count, founding year and LinkedIn page. All fields adds the three extra fields: industry, ticker and revenue.
Final benchmark
The evaluation harness and a public half of the dataset are open source, so any company data provider can be scored on the same companies. Browse the code and data at firecrawl/benchmark-company-enrichment.
Dataset
200 companies scored on the full set. A 102-company public sample, about half of each group, ships for reproduction.
- US public companies
150 companies drawn at random from NYSE and Nasdaq SEC filers, in four size bands from large caps to micro caps. Their answer key comes from SEC filings and their own websites.
Example answer key“amgen.com: Thousand Oaks, California · Pharmaceutical Manufacturing · 31,500 employees · founded 1980 · linkedin.com/company/amgen · AMGN · $36.8 billion revenue”- US private companies
50 companies drawn at random from employers in the Department of Labor's PERM disclosures whose employer ID also files a Form 5500 pension plan, in four size bands. Their answer key comes from those federal filings and their own websites.
Example answer key“providertrust.com: Nashville, Tennessee · Software Development · 116 to 120 employees · founded 2010 · linkedin.com/company/providertrust”
Metrics
How each of the seven fields is scored, and how many of the 200 companies it is graded on. Fields marked [extra] are shown but left out of the headline score.
Headquarters (HQ)
The principal executive office: the newest SEC filing's cover for public companies, federal labor filings for private ones. The same city, or another city in the same US metro area, is correct. Graded on 200 of 200 companies.
Employee count (Employees)
Total headcount from the latest annual report, or from federal payroll and pension filings for private companies. A count within 1.2x of the filed range scores 1, within 2x scores 0.5. Graded on 187 of 200 companies.
Founding year (Founded)
The year the business started; a holding company's or IPO vehicle's later incorporation year does not count. Within one year is correct. Graded on 139 of 200 companies.
LinkedIn page (LinkedIn)
The company's own LinkedIn page, as linked from its website. Any address of the same page counts. Graded on 156 of 200 companies.
Industry[extra]
A label from LinkedIn's industry list. Before any provider was called, two AI models picked the best-fit labels from each company's own description; a best fit scores 1, a broader or secondary label 0.75. A label outside those lists goes to five AI judges, so industry is an extra field. Graded on 199 of 200 companies.
Stock ticker (Ticker)[extra]
Any of the company's common-stock tickers. Public companies only. Graded on 150 of 200 companies.
Revenue[extra]
The latest fiscal year's revenue from SEC filings, within 5%. Public companies only. Graded on 145 of 200 companies.
Frequently asked questions
Changelog
Dated snapshots of this benchmark and what changed in each run.
- Oct 07, 2026People Data Labs is now available in Firecrawl Alexandria. Corrected how the repository README describes Ocean.io's employee count.
- Oct 02, 2026First published run. 6 providers on 200 US companies (150 public, 50 private), scored on four common fields and three extra fields: industry, ticker and revenue. Answer-key corrections made after the provider calls are listed in the repository.
Each snapshot is published from a pinned run in firecrawl/benchmark-company-enrichment.
Related
The product behind this benchmark, and where to read more.
- Firecrawl Alexandriaalexandria.firecrawl.dev