How to Build a B2B Lead List: 5-Step Playbook (2026)

To build a B2B lead list, run five steps in order: define your ICP criteria, choose a data source, enrich each record with contact details, verify the emails, and organize everything under one clean schema. Done right, the output is a deduplicated, personalization-ready sheet or CRM segment you can hand straight to outreach.

Most lead lists fail for boring reasons: no clear definition of a good fit, stale data, unverified emails, and a spreadsheet where every row has slightly different columns. This playbook fixes all four with a repeatable process and two concrete schemas you can copy today.

Step 1: Define Your Criteria (Write the ICP Down)

An Ideal Customer Profile (ICP) is the filter every prospect must pass. For local and SMB B2B, a workable ICP has these dimensions:

• Geography: the cities or radius you serve.
• Category: the business types you sell to.
• Size signal: a proxy you can actually observe - review count, headcount range, number of locations.
• Quality signal: rating floor, "has a website", "operational" status.
• Exclusions: franchises, competitors, existing customers, anyone already in your CRM.

Write it as a single sentence you could hand to a stranger: "Operational independent dental practices within 25 km of Birmingham, with a website and 20+ Google reviews, excluding chains." That sentence becomes your source query and your QA checklist.

Step 2: Source the Businesses

Now turn the ICP into a list of real businesses. You have the three options covered in how to find businesses to contact: manual Google Maps research, a bought directory list, or a business data API. For anything you plan to rebuild regularly, an API wins because the query IS your ICP.

If you are still choosing that source, the business data API guide and the Google Places API vs scrapers comparison cover the trade-offs. With biz collect, the sourcing step is one request that mirrors the ICP sentence:

{
  "location": "Birmingham, UK",
  "keywords": ["dental practice", "dentist"],
  "radius_km": 25,
  "scrape_emails": true,
  "coverage": "exhaustive"
}

coverage: "exhaustive" tiles the radius into a grid to beat Google's 60-results-per-query ceiling, so you approach full coverage of the area instead of the first 60 hits. You get back a businesses array where each record already carries name, address, phone, website, rating, review count, category, hours, and (because scrape_emails is true) emails pulled from each business's own site.

Step 3: Enrich With Contact Details

Discovery gives you the business; enrichment gives you the human path in. This is where map data alone falls short - Google Places returns no emails, so the address has to come from each business's website.

biz collect does this enrichment inside the same job. For every business it visits the site and returns:

• emails - deduped, ranked best-contact-first (on-domain mailboxes above free-provider, above off-domain).
• email_details - per address: a confidence tier, whether the domain has MX records, an is_role_account flag, and a resolved contact (the named person, their title, and a salutation) when the site names one.
• people - the full roster of named people the site published (Impressum / team / about pages), even ones not tied to a scraped email.

That contact object is what lets you write "Hi Dr. Weber" instead of "Hi there" - the difference between a real B2B touch and spray-and-pray. If you would rather enrich a list you already own from raw websites, the extract emails from any website guide covers that path.

Step 4: Verify Before You Send

Sending to unverified addresses burns your domain reputation and inflates bounce rates. Verification has two layers:

1. Syntax + MX (free, automatic in good sources). Is the address well-formed, and does its domain actually accept mail? biz collect does this on every scraped email and exposes it as the confidence tier and has_mx flag - no extra step or cost.
2. Mailbox-level checks (optional, third-party). Dedicated verification tools attempt deeper checks. Note the honest limit: no responsible tool guarantees a live mailbox without risky SMTP probing, so treat "high confidence" as "very likely deliverable", not "guaranteed".

A practical rule: send to high confidence first, hold low/unknown for a warm-up or a manual look, and drop role/no-reply addresses from personalized sequences.

Step 5: Organize Under One Schema

The final step is what separates a lead list from a lead mess: every row shares the same columns, IDs are stable, and duplicates are gone. Here is a lead-list schema you can adopt directly.

Lead list schema (one row per business)

Dedupe rules (apply in order)

1. Exact website host match (strip www., lowercase).
2. Exact normalized business_name + city match.
3. Same primary_email across two rows.

Keep the row with the higher email_confidence; merge phone/social from the discarded one.

Because biz collect returns stable JSON and a CSV/XLSX export with fixed columns, mapping into this schema is a straight field-to-column copy - no HTML parsing, no per-source cleanup. Teams that want this to run on a schedule can wire it into n8n, Make, or an AI agent pipeline.

A Worked Example

Say you want dental practices around Birmingham. You run the Step 2 request, get 84 businesses back, of which 61 have a website and 47 yield at least one high or medium confidence email. After deduping on website host you are at 44 unique leads. You map them into the schema above, tag the 31 with a resolved contact_name as "personalize", and route the 13 role-only inboxes to a lighter, non-personalized touch. That is a clean, send-ready B2B lead list built in one afternoon - and rebuildable next quarter by re-running one request.

The Bottom Line

A B2B lead list is a small data product: defined by an ICP, sourced fresh, enriched with real contacts, verified for deliverability, and organized under one schema with dedupe rules. Nail the schema and the process, and every future campaign starts from a clean, repeatable base instead of a blank spreadsheet.