How to Outsource B2B Lead Generation (and Audit the Agency's List)

A B2B lead generation agency sells you a list, meetings, or both. The risk is that you cannot see what they sourced, so you pay for a black box. This guide gives you a brief an agency cannot fudge and a way to check their output yourself.

TL;DR

• Size your market first with a free count, so the agency's quote is checked against a real number.
• Pilot on 100 rows and spot-check 25 of them through an API. Total cost is about 125 LeadOcean records.
• Setup takes about 2 hours. The pilot runs 2 to 4 weeks. Keep the agency only if the pilot passes your own thresholds.

Prerequisites

• A LeadOcean account on the Free plan (1,000 records for life, no card, 5 requests a second). Sign up at app.leadocean.io.
• An API key stored as $LEADOCEAN_API_KEY.
• curl and jq, or the LeadOcean MCP server at https://api.leadocean.io/mcp in Claude.
• One sentence describing your buyer: title, department, company size, country.

Step 1: Size the market before you talk to anyone

Run a free count so you know how many buyers exist. An agency that promises 5,000 qualified leads from a market of 54 is selling you something else.

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", "C-Team"],
    "jobFunction": ["Sales & Business Development"],
    "country": ["US"],
    "employeeRange": ["51-200"],
    "industry": ["Computer Software"],
    "limit": 1
  }' | jq '.meta'

count=true is a query parameter. It spends no records. Note that meta.total is capped at 100,000.

Step 2: Write the brief as filters, not adjectives

"Decision makers at growing SaaS companies" cannot be audited. A filter set can. Write the brief as the exact fields you will check later.

{
  "jobLevel": ["VP", "C-Team"],
  "jobFunction": ["Sales & Business Development"],
  "country": ["US"],
  "employeeRange": ["51-200"],
  "industry": ["Computer Software"],
  "emailStatus": ["verified", "catch_all_valid"]
}

Enum values are exact strings. Resolve them with the free leadocean_list_enum_values MCP tool before you send the brief. Then add the delivery terms: rows per week, the email status you will accept, and a rule that every row carries a source and a date.

Step 3: Pull a control sample of 100 rows

Fetch 100 people from the same filters yourself. This is your yardstick. If the agency's sample looks nothing like it, ask why.

curl -s -X POST "https://api.leadocean.io/v1/people/search" \
  -H "x-api-key: $LEADOCEAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jobLevel": ["VP", "C-Team"],
    "jobFunction": ["Sales & Business Development"],
    "country": ["US"],
    "employeeRange": ["51-200"],
    "industry": ["Computer Software"],
    "limit": 100
  }' > control.json

This spends up to 100 records. Page with cursor if you need more. One search pages to 10,000 rows at most.

Step 4: Spot-check 25 rows from the agency's pilot list

Take 25 random rows from the agency's file and enrich each by linkedin_url or email. Compare title, company and email status to what they claimed.

curl -s -X POST "https://api.leadocean.io/v1/people/enrich" \
  -H "x-api-key: $LEADOCEAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"linkedin_url": "https://www.linkedin.com/in/jane-doe-example"}'

Enrich takes exactly one of linkedin_url, email, phone or person_id. A person we hold nothing for is a 404 and costs nothing. Set pass thresholds before you look: for example, 90% of titles match and 95% of emails are not invalid.

Step 5: Score the pilot and decide

Count matches in a spreadsheet. Check that title and company still match, that the email status is what they promised, and that no row is on your do-not-contact list. Read email_status before sending. Bounced emails are not refunded.

If the agency fails, you lose 2 to 4 weeks, not a year-long retainer.

Step 6: Put the pilot terms in writing

Turn the thresholds into a contract schedule. Agencies sell three models: a list only, appointments set, or a full outbound program. Each needs different terms, and you can audit all three with the same checks.

pilot:
  duration_weeks: 3
  rows_delivered: 100
  filters: brief.json            # the Step 2 filter set
  email_status_allowed: [verified, catch_all_valid]
  per_row_fields: [title, company_domain, email, source, fetched_date]
acceptance:
  title_company_match: 0.90      # of the 25-row spot check
  email_not_invalid: 0.95
  duplicates_vs_do_not_contact: 0
remedy:
  below_threshold: replace_rows_free
  second_failure: terminate_no_fee

For a list-only deal, pay per accepted row. For appointments, define a qualified meeting in the contract: who attended, their title, and the date. For a full program, ask who owns the sending domains, because you will want them if you leave.

Ask where each row came from. A fair answer names a data provider or a first-party source. "Proprietary database" with no field-level detail is a reason to run the Step 4 check on 50 rows, not 25.

Step 7: Decide on a retainer or a cheaper path

If the pilot passes, move to a month-to-month retainer with the same acceptance rules. If it fails, compare the cost with doing it yourself. A Solo plan is $97 a month and an Agency plan is $497 a month (see pricing). Paid plans have a fair-use pace, not credits.

You still need sending infrastructure and someone to write and reply. That is the part an agency really sells. If you already have it, a data API and one SDR may beat a retainer.

What you get

The Step 1 filters, counted with the free LeadOcean MCP leadocean_count_leads tool on 2026-10-21. It counted only addresses worth sending to (verified, catch_all_valid or catch_all), so the total is a mailable audience, not everyone.

{
  "total": 54,
  "totalIsExact": true,
  "totalAll": 205,
  "distinctCompanies": 30,
  "credits": 0
}

54 mailable people at 30 companies, out of 205 with no email condition. A market this size does not need an agency to "generate" thousands of leads.

A single search row has this shape (response shape from the API reference, placeholder values):

{
  "success": true,
  "data": [
    {
      "profile_data": {
        "profile_full_name": "Jane Doe",
        "profile_url": "https://www.linkedin.com/in/jane-doe-example"
      },
      "contact_data": {
        "has_email": true,
        "email_status": "verified",
        "contact_current_experiences": [
          { "job_title": "VP Sales", "company_name": "Acme", "company_domain": "acme.com" }
        ]
      },
      "meta": { "source_ids": { "person_id": "p_0000000" }, "fetched_at": "2026-10-01" }
    }
  ],
  "meta": { "count": 1, "nextCursor": "c_0000000" }
}

Troubleshooting

FAQ