DEV Community

Jack Dang
Jack Dang

Posted on

n8n workflow: enrich a Google Sheet of domains with company data (name, size, funding), plus a 99-domain benchmark

What you get: an n8n workflow that takes a column of company domains in Google Sheets and fills in name, description, industry, HQ, employee range and funding on the same rows. On a hand-picked list of 99 domains it finished in about 5 minutes 45 seconds with no manual lookups. A second template builds a company list from a keyword instead.

Why use it

  • No copy-pasting between tabs. Put domains in column A, click Execute, and every row is enriched in place.
  • Nothing to code. Import workflow.json into n8n, connect Google Sheets, add one API credential.
  • You can trust what a row means. Each row ends with a status: found, found_live_only, not_found, skipped_duplicate or error. Failed lookups are never written as blank rows.
  • Safe to re-run and cheap to run. Rows that already have a status are skipped. You pay 1 credit per company found; misses and errors are free.
  • Works in the other direction too. Template 2 turns a keyword, countries and a size range into a list of matching companies, skipping ones already in your sheet.

Disclosure: I run ReplyNodes, the company database these templates call, and I wrote the templates. They are MIT licensed, but they need a ReplyNodes API key. New accounts get 500 free credits; after that it is paid (see replynodes.com for pricing). The templates are not in the n8n template library yet, so import them from the repo: https://github.com/replynodes/n8n-company-research

At a glance

Template 1: enrich by domain Template 2: ICP list builder
Input domains in a Google Sheet keyword, countries, size range
Output company fields added to the same rows matching companies appended to a sheet
Cost 1 credit per company found 1 credit per page of up to 10 companies
Result I measured 99 domains: 93 found, 4 website-only, 2 not found, 0 errors 1 page, 6 companies, 1 credit
Tested on n8n 2.35.7 and n8n Cloud, real Google Sheet same

Last updated 2026-11-20.

How do I enrich a Google Sheet of domains in n8n?

Put domains in column A, click Execute, and the same rows get filled in:

The workflow filling a Google Sheet of domains with company data

Columns per row: status, source, name, description, industries, hq_country, hq_region, employees_min, employees_max, funding_total_usd, latest_round_type, latest_round_date, data_loaded_at, checked_at and error.

  • Input. Full URLs are normalized (https://www.Stripe.com/pricing becomes stripe.com). Invalid domains get an error status. Duplicates in the same run are skipped and marked.
  • Statuses. found_live_only means the company is not in the database and only the name and description were read from the live website, so do not expect size or funding. A hit counts only if its domain is exactly the one you asked for.
  • Re-running. Only rows with an empty status are processed, so you can add rows and run again. To re-check a row, clear its status cell.
  • Pace. One lookup at a time, about one per second (that is what the API allows), with a single retry on 429, 503 or 504. A bad row never stops the run.
  • Credits. 1 credit per lookup that finds a company. Lookups that find nothing, and errors, are free. The website fallback uses a keyless endpoint with its own public daily limit.

How do I build a list of target companies from a keyword in n8n?

Template 2 appends matching companies to a sheet and skips companies already in it. In the Settings node you set:

  • keyword, one term or a quoted phrase, for example carbon accounting software
  • countries, for example US
  • employeesMin and employeesMax, for example 11 and 200
  • industries (optional) and maxPages (default 5, capped at 10)

Cost is 1 credit per page of up to 10 companies; empty pages and errors are free, and maxPages caps the spend. On a real run with the settings above and maxPages 1 it fetched one page, added 6 companies, stopped with "no more results" and used 1 credit.

Search is by keyword, not by meaning: "top 10 fintech startups in Berlin" will not work, a plain term plus filters will. Avoid OR and -exclusion in the keyword, those queries often hit the API's 5-second limit. A search returns at most 1,000 results.

How well does it work? A 99-domain benchmark

I ran template 1 on 99 domains: 97 well-known SaaS, developer-tool and climate-tech companies, plus 2 made-up domains to check the failure path. Default settings, on n8n Cloud, writing to a real Google Sheet.

Result Rows
found (in the database) 93
found_live_only (website metadata only) 4
not_found (the 2 made-up domains) 2
error 0
  • Of the 93 database hits, all 93 had an employee range, 82 had a description and 78 had funding data. So funding was missing on 15 of the 93.
  • It used 93 credits (one per found company; the live-only, not-found and error rows were free) and took about 5 minutes 45 seconds, roughly 3.5 seconds per row end to end.
  • No retries were needed.

What are the limits?

The list is hand-picked and popular, so the hit rate on obscure domains will be lower. A smaller earlier check on 20 well-known domains returned the right company for 16, with calendly.com, typeform.com and firecrawl.dev missing (the website fallback fills those in). Please do not read 93 of 97 as a promise for your own list.

The data comes from third-party datasets and it is not perfect. While spot-checking I found at least two wrong matches: gumroad.com came back as a record named AirPitch, and notion.so came back as a different company. Some industry labels are odd (Stripe is labelled "facilities services"). Lookup is by exact domain only, not by company name. Every row carries data_loaded_at, which shows when the record was last loaded. Treat the output as a good first pass and spot-check anything you are going to act on. For template 2, keyword results can include loosely related companies, so read the description column before using a list for outreach.

This is a lightweight option for these fields, not a replacement for a full enrichment platform.

How do I set it up?

  1. Create a Google Sheet and paste the header row from the README of the template you use (template 2 adds query and page columns).
  2. Download workflow.json from the repo and import it (Workflows, Import from File).
  3. Open the Google Sheets nodes (template 1: Read domains and Write row; template 2: Read existing rows and Add to sheet), connect your Google account and pick the same spreadsheet and tab in both.
  4. Create a Header Auth credential called ReplyNodes API with header name Authorization and value Bearer <your key>, and select it in the HTTP Request nodes (template 1: both Search company nodes; template 2: Search page).
  5. Template 2 only: open Settings and describe your target. Then click Execute workflow.

One n8n quirk worth knowing: every time you pick a document or sheet in a Google Sheets node, n8n resets the column mapping. After choosing the sheet, set Mapping Column Mode back to Map Automatically (and, in Write row of template 1, set Column to match on to row_number). The READMEs say so, but it is easy to miss.

A wrong API key shows up in the error column as an HTTP 401 message ending with "(check the ReplyNodes API credential)", not as a silent empty sheet.

FAQ

Does it need coding? No. You import the workflow JSON and connect two credentials: Google Sheets and a Header Auth credential for the API.

How much does it cost? Template 1 costs 1 credit per company found. Template 2 costs 1 credit per page of up to 10 companies. New accounts get 500 free credits, then it is paid.

What if a company is not in the database? The row is marked found_live_only (name and description from the website) or not_found, and it costs nothing.

Can I look up by company name? No, template 1 is an exact-domain lookup.

What does the X-Source header do? Requests carry n8n-company-research/enrich-by-domain or n8n-company-research/icp-list-builder so I can count template usage. Only the domains you submit (template 1) or your search settings (template 2) are sent. To remove it, clear the sourceHeader value in the Settings node and the header in the HTTP Request nodes.

Links

If you try them on a real list, I would like to hear what the hit rate looks like on yours, and which columns you wish were there.

Top comments (0)