Lead List Deduplicator & Normalizer - CRM-Ready Leads, Not Messy Dumps

Turn messy scraped B2B lead lists into canonical, CRM-ready records - not duplicate-filled dumps.

This Actor takes inline JSON records or an Apify dataset ID, normalizes common lead fields, groups duplicates, and outputs one canonical row per lead/company cluster with confidence scores, match reasons, source row IDs, and warnings. Use it after Google Maps scrapers, directory scrapers, website contact scrapers, exhibitor-list scrapers, Apollo-style lead exports, or any workflow where several sources produce overlapping leads.

✅ What you get / ❌ what this isn't

✅ This Actor gives you	❌ This Actor is not
One canonical row per company/contact cluster	Not a black-box cleanup you can't audit
Confidence scores + match reasons per merge	Not a guess - source row IDs are preserved
Normalized company, domain, email, phone	Not a scraper or enrichment tool (it cleans what you give it)
Deterministic, predictable-cost rules	Not a paid LLM that adjudicates every row

🔎 Why use this Actor

• Merge overlapping exports from multiple scrapers.
• Remove duplicate companies, domains, emails, and phone numbers before CRM import.
• Normalize company names, domains, emails, and phones.
• Keep source row IDs so every merge is auditable.
• Get confidence scores and match reasons instead of a black-box cleanup.
• Use deterministic rules first, so costs stay predictable.
• No browser, proxies, or external enrichment APIs.

👥 Who it's for

Anyone importing scraped or exported B2B leads into a CRM. Common jobs:

• Merge lead lists from several Apify scrapers.
• Clean a CSV before importing into HubSpot, Pipedrive, Salesforce, Clay, Instantly, Smartlead, or Airtable.
• Remove duplicate outreach targets before spending credits on email verification or enrichment.
• Create a canonical company list from multiple scraped directories.
• Audit which rows were merged and why.

⚙️ How to deduplicate a lead list

1. Open the Actor on Apify.
2. Paste your records (inline JSON) or provide an Apify datasetId.
3. Pick a dedupMode: conservative, balanced, or aggressive.
4. Click Start.
5. Open the Canonical view for CRM-ready rows, or Duplicate clusters to audit merges.
6. Download CSV/JSON/Excel or pull from the Apify API.

If no input is provided, the Actor runs with sample records so you can test the output immediately.

📥 Input

{
  "dedupMode": "balanced",
  "records": [
    {
      "id": "1",
      "company": "Acme Inc",
      "website": "https://www.acme.com",
      "email": "sales@acme.com"
    },
    {
      "id": "2",
      "companyName": "ACME LLC",
      "domain": "acme.com",
      "phone": "(415) 555-2671"
    }
  ]
}

You can also provide an Apify datasetId instead of inline records.

Deduplication modes

Mode	Best for	Behavior
conservative	Avoiding false merges	Requires exact email, phone, or domain match
balanced	Most lead lists	Exact email/phone/domain plus strong company-name similarity
aggressive	Very messy lists	Looser company-name matching; review warnings before importing

📤 Output

{
  "recordType": "canonicalLead",
  "clusterId": "cluster_0001",
  "clusterSize": 2,
  "mergeDecision": "merged",
  "mergeConfidence": 0.9,
  "matchReasons": ["same_domain", "similar_company"],
  "sourceRowIds": ["1", "2"],
  "canonicalCompanyName": "Acme Inc",
  "normalizedCompanyName": "acme",
  "normalizedDomain": "acme.com",
  "normalizedEmail": "sales@acme.com",
  "normalizedPhone": "4155552671",
  "warnings": []
}

Dataset views

View	Best for
Canonical	CRM-ready rows after deduplication
Duplicate clusters	Auditing source rows, match reasons, and confidence

Output fields

Field	Meaning
clusterId	Stable cluster identifier for the canonical row
clusterSize	Number of source rows merged into the canonical row
mergeDecision	unique, merged, or ambiguous
mergeConfidence	Confidence score from 0 to 1
matchReasons	Why records matched (same_email, same_domain, similar_company)
sourceRowIds	Original row IDs or indexes used in the merge
normalizedDomain	Clean domain value such as acme.com
warnings	Flags such as low_confidence_merge or missing_domain_or_email

💵 How much does it cost?

You pay per cleaned output row plus Apify platform usage. Because the engine is deterministic (no browser, no proxies, no external APIs), cost is predictable and scales with input size. Each run processes up to 5,000 input records; split larger datasets across multiple runs.

🔁 Run it on the Apify platform

Chain it after any Apify scraper via the API, schedule recurring cleanups, export CSV/JSON/Excel, or wire it into Make, Zapier, or webhooks ahead of your CRM import.

⚠️ Limits and caveats

• This MVP uses deterministic rules and fuzzy string similarity, not paid LLM adjudication.
• Review ambiguous rows before importing them into a CRM.
• Email/phone/domain normalization is conservative and may not cover every country-specific format.
• The Actor does not scrape or enrich missing contact data; it cleans the records you provide.
• It does not verify email deliverability or MX records in this version.
• Runs are capped at 5,000 input records while the engine is optimized for larger files.

🧩 Related Actors

• Website Contact Extractor - find the emails first, then dedupe them here.
• LinkedIn Ads Library Scraper - build the advertiser list this cleans.

❓ FAQ

Does it scrape leads? No. It cleans and dedupes the records you provide (inline or via a dataset ID).

Can it pull from another Actor's output? Yes - pass that run's datasetId as input.

Which mode should I use? balanced for most lists; conservative to avoid false merges; aggressive only for very messy data (then review warnings).

🛠️ Support

If a run fails or a field is missing, open an Actor issue with the run URL, the input you used, and the field or behavior you expected.