Companies House SIC Code Scraper

This Apify actor scrapes Companies House to find companies with similar SIC codes and formats the results for use as Google Ads negative keywords.

Features

• Searches Companies House for a seed company and extracts its SIC codes
• Finds other companies with the same SIC codes
• Filters results by address (optional)
• Cleans and formats company names for use as negative keywords
• Identifies minimum negators based on dictionary words
• Detects blocked keywords

Input

The actor accepts the following inputs:

• seedCompany (required): The name of the company to search for
• addressFilter (optional): Filter results by address (e.g., "London", "England")
• dictionary (optional): List of dictionary words for finding minimum negators
• keywords (optional): List of keywords to check against minimum negators
• locations (optional): List of location names to remove from company names

Output

The actor produces two outputs:

1. Main dataset: Formatted results with:

   • rawNameTitleCase: Original company name in title case
   • cleanedName: Cleaned company name
   • minimumNegator: The minimum phrase needed as a negative keyword
   • blockedKeywords: Keywords that would be blocked by this negator

2. Key-value store (raw-companies): Raw company data including:

   • Company name, number, status, type
   • Address and incorporation info
   • SIC codes

Example

{
  "seedCompany": "SOLOWASTE LTD",
  "addressFilter": "London",
  "dictionary": ["the", "and", "of", "in", "to", "for"],
  "keywords": ["waste management", "waste disposal", "recycling"],
  "locations": ["London", "Manchester", "Birmingham"]
}

How it works

1. The actor searches Companies House for the seed company
2. Extracts all SIC codes associated with that company
3. For each SIC code, searches for other companies with the same code
4. Filters results by address if specified
5. Cleans company names by removing corporate identifiers and locations
6. Finds the minimum negator based on dictionary words
7. Identifies which keywords would be blocked by each negator