What we ran
- Brands: 10 well-known shop brands, picked by us.
- Names: the generator behind the lookalike domain finder: dropped, doubled, swapped and neighbour-key letters, look-alike characters, hyphens, 24 other endings, and added words like shop and store.
- Check: one DNS query per name (Cloudflare DNS over HTTPS), from our test box. “Exists” means the name is not NXDOMAIN. We loaded no pages. See lookalike domains examples for pages.
What we found
1,347 names, 614 exist (46%), 539 of those resolve to an address (40%).
| kind of typo | names | exist | share |
|---|---|---|---|
| letter dropped | 74 | 65 | 88% |
| other ending | 240 | 192 | 80% |
| vowel swapped | 83 | 43 | 52% |
| letter doubled | 74 | 38 | 51% |
| letters swapped | 64 | 29 | 45% |
| look-alike character (m for rn) | 34 | 12 | 35% |
| neighbour key | 379 | 133 | 35% |
| added word (shop, store) | 280 | 93 | 33% |
| hyphen | 66 | 9 | 14% |
| Unicode look-alike | 53 | 0 | 0% |
Every brand had .net, .org, .co, .eu, .de, .fr, .ca and .nl taken. Example names that exist: albirds.com, gymshrak.com, gym-shark.com, allbirds.shop, shopallbirds.com.
Existing is not squatting. Names near a short brand are often other companies: chewy.com is one letter from chew. The page decides, so this sweep is the cheap first cut before a page check. Rule on the term: typosquatting.
How you’d use it
- Brand protection: pre-filter. Register the endings that matter, watch the rest weekly, load pages only for names that resolve.
- Registrar or regulator: the same sweep over a protected name list gives a dated table per brand.
- Data team: the full name list per brand is a custom feed, with a diff when new names appear.