Lesson 2.3.2
Merge, import, and deduplicate without creating ghosts
What this costs
Free, and it saves credits by removing rows you would otherwise enrich twice.
Before you start
Two or more projects covering an overlapping market.
Two builders on the same market produce a better list than either alone, but only once the overlap is resolved. Deduplicating on company name alone will quietly keep three versions of the same company and delete two that were genuinely different.
Deduplicate on the domain, not the name
Company names are written differently by every source. Ltd, Limited, GmbH, the trading name, the legal entity, a typo. Domains are canonical. Resolve websites first, then deduplicate on the domain, and the whole problem mostly disappears.
- 1
Resolve domains on both lists first
Website Finder on the gaps, before you merge. Merging first and resolving after means you deduplicate against a field half your rows do not have.
- 2
Merge into one project
Bring the sources together. Keep a column recording which source each row came from, because that is a quality signal in itself.
- 3
Preview the deduplication
Read what it proposes to collapse before it does it. This is where you catch two genuinely different subsidiaries sharing a parent domain.
- 4
Decide which value survives a merge
When two rows disagree on a field, one has to win. Prefer the source with the better provenance, and keep the other value in a secondary column if the difference matters.
- 5
Check for companies you already have elsewhere
Before enriching, check the merged list against your other projects. Rows you already own do not need paying for twice.
Deduplicated on company name
- What you see
- Duplicates survive and distinct companies get merged
- The fix
- Resolve domains first and deduplicate on those.
Merged before resolving websites
- What you see
- Dedup runs against a field half the rows lack
- The fix
- Website Finder first, then merge, then dedup.
Lost the source attribution
- What you see
- No way to tell which builder produced better rows
- The fix
- Add a source column before merging. It is free and it is the only way to improve your sourcing over time.
Check yourself
0 of 2 answered1.What should you deduplicate on?
2.Why record which source each row came from before merging?
Related and next
- Help: Merge multiple projects into one
- Help: Check for existing companies
- Next: duplicate a workflow without its data
Recipes that use this skill
This lesson is written and complete. Its screen recording is still to come, and the written walkthrough is designed to stand on its own without one.