How to Benchmark Data Enrichment Platforms (Your Own 100-Contact Test)

Skip the 12-vendor leaderboards, including ours: we have not published one, and a public score cannot tell you how a vendor does on your market. Run a blind test on 100 contacts you can verify. This guide shows the method with LeadOcean as the first vendor and one scoring sheet that works for any other.

TL;DR

• Build a list of 100 people you can verify, run each vendor on it, and score hit rate, email status and phone coverage. Plan on about 90 minutes.
• On LeadOcean the test spends at most 100 of your 1,000 free records. Misses are free 404s, and every sizing count is free.
• Repeat the same script per vendor and compare on your list. Do not compare vendor totals.

Prerequisites

• A free LeadOcean key from app.leadocean.io, exported as $LEADOCEAN_API_KEY. No card.
• A free or trial account at each vendor you want to test, with their API key in its own variable.
• 100 contacts with LinkedIn URLs, taken from your CRM and closed-won deals. Not scraped from a vendor.
• curl, jq and a spreadsheet.

Step 1: Size your market before you test anything

Start with a free count so you know whether the vendor holds your segment at all. A count spends no records, however often you run it.

curl -s "https://api.leadocean.io/v1/people/search?count=true" \
  -H "x-api-key: $LEADOCEAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jobLevel":["VP"],"jobFunction":["Sales & Business Development"],"country":["DE"],"limit":1}'

count=true goes in the query string, not the body. The response has meta.total (capped at 100,000) and an empty data array.

Step 2: Build the test list

Take 100 rows with one LinkedIn URL each. Mix 70 current customers or known leads with 30 hard cases: small firms, non-US contacts, people who changed jobs lately. Hard cases are where vendors differ.

linkedin_url,known_title,known_company,known_country
https://www.linkedin.com/in/jane-doe-example,VP Sales,Acme,DE

Write the truth you already know in the other columns. You will score each vendor against it.

Step 3: Run LeadOcean on the list

Enrich takes exactly one of linkedin_url, email, phone or person_id. Free keys run 5 requests a second, so pace the loop.

mkdir -p out
tail -n +2 list.csv | cut -d, -f1 | while read url; do
  id=$(echo "$url" | md5)
  curl -s -o "out/$id.json" -w "%{http_code}\n" -X POST "https://api.leadocean.io/v1/people/enrich" \
    -H "x-api-key: $LEADOCEAN_API_KEY" \
    -H "Content-Type: application/json" \
    -d "{\"linkedin_url\":\"$url\",\"reveal_email\":true}" > "out/$id.code"
  sleep 0.25
done

On Linux use md5sum | cut -c1-12 instead of md5. A 404 means we hold nobody for that URL and it costs nothing. A 200 costs one record.

Step 4: Score every vendor the same way

Count a hit only when the returned title and company match your truth column. A profile returned for the wrong person is a miss, not a hit.

for c in out/*.code; do
  if [ "$(cat "$c")" = "200" ]; then
    jq -r '.data.contact_data | [.contact_current_experiences[0].job_title,
      .contact_current_experiences[0].company_name,
      .email_status, ((.contact_phones // []) | length > 0)] | @tsv' "${c%.code}.json"
  fi
done | tee hits.tsv
wc -l hits.tsv

Put these columns in the sheet for each vendor:

LeadOcean returns an email status with each address, among them verified, catch_all_valid, catch_all, risky and invalid. Ask every other vendor for the equivalent and score it the same. If a vendor gives no status, treat every address as unverified.

Step 5: Send a sample of emails through a checker

Presence is not accuracy. Take 30 addresses per vendor and run them through one independent verifier, then compare bounce rates. This is the only step that tells you whether the data works.

for c in out/*.code; do
  [ "$(cat "$c")" = "200" ] && jq -r '.data.contact_data.contact_emails[0].email // empty' "${c%.code}.json"
done | head -30 > sample_emails.txt
wc -l sample_emails.txt

Addresses sit in contact_data.contact_emails[], best first, each with its own email, type and status (enrich docs).

Use the checker you already pay for. Run the identical file for each vendor on the same day, because deliverability drifts.

Step 6: Price the winner at your real volume

Cost per hit depends on volume and on whether misses are billed. Read each vendor's pricing page on the day you test, and note whether a lookup that finds nothing is charged.

LeadOcean has no credits on paid plans: Solo is $97 a month, Agency is $497 a month and Platform is $1,997 a month, each with a fair-use pace. Free is 1,000 records for life. A search page of 25 is 25 records, and an enrichment that finds nothing costs none.

For a vendor with credits, multiply only a rate you can read on the page. If the page shows no price at your volume, record "No public price at volume" rather than extrapolating.

Step 7: Read the results with a rule set

Decide your pass marks before you look at the numbers. Otherwise the vendor you already like wins.

• Set a minimum hit rate on your 70 core contacts before you start, and judge the 30 hard cases separately.
• Rank vendors on verified-email share first. A high hit rate with unverified addresses means bounces.
• Reject any vendor whose wrong-person rate is above a few percent. A confident wrong profile costs more than a miss.
• Rerun the winner and the runner-up with a second, fresh list of 100 a week later. If the order flips, the gap is noise.

Keep the sheet. Rerun the same 100 contacts every quarter, since vendors change their data and their pricing.

What you get

The sizing call from Step 1, run through the free leadocean_count_leads MCP tool on 2026-10-21:

{
  "filters": {"jobLevel": ["VP"], "jobFunction": ["Sales & Business Development"], "country": ["DE"]},
  "total": 1395,
  "totalIsExact": true,
  "totalAll": 3893,
  "distinctCompanies": 905,
  "credits": 0
}

The total of 1,395 counts only people whose best email is verified, catch_all_valid or catch_all. Without an email condition the same filters match 3,893. That gap is the first number to compare across vendors.

The enrich response carries the current role in contact_data.contact_current_experiences[], addresses in contact_data.contact_emails[], phones in contact_data.contact_phones[].phone, and career history in resume_data when held. Response shape from the API reference, with placeholder values:

{
  "success": true,
  "data": {
    "contact_data": {
      "email_status": "verified",
      "contact_emails": [{"email": "jane.doe@acme.com", "type": "work", "status": "verified"}],
      "contact_phones": [{"phone": "+1 415 555 0133", "type": "mobile"}],
      "contact_current_experiences": [{"job_title": "VP Sales", "company_name": "Acme"}]
    },
    "resume_data": {"experiences": [], "educations": []}
  }
}

Troubleshooting

FAQ