People research is where bad OSINT habits show up fastest. The goal is not to collect every possible clue. The goal is to decide whether several public signals point to the same person strongly enough to keep going.
Recommended sequence
- Start with the strongest clue you actually have: name plus employer, email, username, or phone number.
- Use fast enrichment tools such as Epieos for initial signal, then check whether the result fits public context rather than trusting the first output.
- If the case touches work identity, use Hunter or company context to test whether a professional link is plausible.
- Archive relevant pages and use historical or breach-search tools only when the lead is worth the extra verification burden.
Where this goes wrong
Name collisions, recycled usernames, and public profile fragments can make unrelated people look connected. One matching clue is rarely enough.
Before you publish
Write down what is directly observed, what is inferred, and what remains unconfirmed. That separation matters most when the subject is a real person.