How to tell same-name candidates apart
Identity checks, and where to stop

01

Why are same-name candidates so easy to mix up?

A name is the least reliable piece of identity there is. It’s nothing unusual for a common name to turn up twice in the same industry and the same city. It works the other way, too: one person’s name can be written completely differently in different places. The CV gives the name in Chinese characters, a paper uses abbreviated pinyin (the standard romanization), an open-source community shows an English username, and the company’s own page uses their English name.

• Same name, different people

  Two people share a name and their records get merged. You end up judging one person by another’s experience, or even contacting the wrong person.

• Same person, different names

  One person’s name is written differently in several places, and nobody spots that it’s the same person. They appear twice on your list and get contacted twice.

Take an example. A paper’s author listed as “Lin Y.” could be the Yue Lin in your database, or another researcher with the family name Lin and a given name starting with Y. Merging on an initial alone is a big risk. Both mistakes hurt. Getting someone’s experience wrong in a recommendation costs you with the client. Contacting the wrong person, or the same person twice, costs you with the candidate.

02

What do you compare in a candidate identity check?

What a headhunter checks is professional identity: whether the records in front of you describe the same person, and whether that person really did this work at these companies. The main things you can compare are these:

• Employers: a full name, a short name, an English name, a subsidiary or the name before an acquisition may all point to the same company.
• Dates: the start and end dates at each company, and whether they line up with the years of papers, patents and projects.
• City: where they work, and whether any move between cities has a plausible reason.
• Field: whether the work follows one consistent thread, or a completely unrelated stint suddenly appears.
• Papers, code and talks: the affiliation on a paper, co-authors, the company named in an open-source project, the title on a conference program.

No single item is enough. A big company may have two people with the same name working in the same field, and being in the same city proves little. Only when several items line up and support each other is the case convincing. Wherever one doesn’t match, it’s worth stopping for a second look.

All of this comes only from the CV the candidate gave you and from professional settings such as company websites, team pages, papers, patents, open-source projects, conference programs and published news coverage.

03

How do you cross-check, and when is it really a match?

1. Pick an anchor

   Take the record you’re most sure of as your baseline, such as the CV the candidate sent you, or the team page on the company’s website.

2. Compare item by item

   Compare each other source with the anchor: company, dates, city, field, co-authors. Note what matches, what doesn’t and what isn’t mentioned, each separately.

3. Check that the timeline holds

   No one should be full-time at two companies at once, and changes in the affiliation on someone’s papers should roughly match when they changed jobs.

4. Count independent sources only

   An article reposted many times still counts as one source. You’re on fairly solid ground only when at least two independent sources agree.

5. If anything conflicts, don’t merge yet

   Even if most things match, one clear conflict is enough to keep the records apart until you’ve worked it out.

In checking consistency, the most useful clues are often the small ones: the same co-author turning up across several papers, the employer on an open-source project page matching the CV, the title on a conference program matching the team page. Details like these rarely all line up by chance. Conversely, if someone’s CV is industrial vision all the way through, but a paper under the same name lists a hospital imaging department, it’s worth asking whether these are two people.

04

When you’re not sure, how do you flag it and confirm it?

A check doesn’t only come out as yes or no. Write down how confident you are, so the next time you or a colleague looks, it’s clear how far the information can be relied on. You can use four levels:

• Confirmed

  Several independent sources agree, or the candidate has confirmed it. You can merge the records into one person.

• Fairly sure

  The main details match, but one key piece of evidence is missing, for instance the dates haven’t been matched up yet. Don’t merge yet, and note what’s missing.

• To confirm

  Only the name matches, or there’s only one source, or something conflicts. Treat it as a lead only, and keep it out of any recommendation.

• Ruled out

  Clearly not a match. Record why, so the two aren’t taken for the same person again later.

Next to each judgment, note what it rests on: which sources, what each one said, what’s still unconfirmed, and when you checked. Don’t write unconfirmed information as if it were certain, and don’t merge someone still unconfirmed just to get a list out faster.

If you really can’t confirm it, the most direct way is to ask the candidate. In your first message you can mention it naturally: “I understand you worked on computer vision at [company]. Have I got that right?” They’ll tell you, and they’ll see you’ve done your homework. You can also ask a mutual contact you trust, who knows them, to help confirm. If the candidate doesn’t reply, leave it unconfirmed, and don’t go looking for other ways to find out.

05

How does HunterBuddy help you tell same-name candidates apart?

The method isn’t hard. The volume is. Everyone on a longlist may be scattered across your CV database, team pages, papers, patents and open-source projects, and checking record by record takes a long time. HunterBuddy, the AI partner for headhunters, takes over the cross-checking and leaves the judgment to you. Here’s the example our homepage uses throughout: Xingchuan Robotics is hiring a Head of Computer Vision, and say Yue Lin has one record each in your CV database, on a team page and in a paper.

• When the same person is scattered across different sources, it checks their identity before merging. It won’t treat two records as one person just because the names match.
• Leads that come from paper authors or patent inventors are flagged as still needing an identity check.
• Anything without a reliable source is flagged for you to confirm. Inferences are kept apart from confirmed facts, and every conclusion names its source.
• CVs you upload and org charts you import go through the same identity checks, and are de-duplicated and merged.
• Once identities are clear, it suggests a suitable way in, such as who could introduce you.

• Merged · 3 sourcesCV database · Team page · Paper

  Company, dates and field all match. Merged into one person.

• Ruled out · NamesakeAnother person with the same name

  Neither city nor field matches. Not merged.

• To confirmTeam size and reporting line

  No reliable public source. You’ll need to ask.

Whether these records belong to the candidate in your database, and whether to merge them, is yours to confirm. Who to contact and how to open the conversation is your decision too. It won’t contact candidates for you. HunterBuddy is currently in development, with an invitation-only trial.

06

How far should a candidate identity check go?

A headhunter’s identity check has one purpose: to confirm that the person you’re about to contact and recommend is the same person who did this work in their professional life. That’s as far as it needs to go. There’s no need to dig further, and you shouldn’t.

• No background checks. A background check is a separate process that happens after the candidate authorizes it, and it isn’t part of finding people.
• No looking up personal information unrelated to work, such as ID numbers or home addresses.
• Identities are matched only from information made public in professional settings, never from private details.
• Use only what the person has made public in professional settings, plus what the candidate has given you.
• If someone keeps something private, or doesn’t reply, that’s their right. Respect the choice.

Some people use different names in different settings, such as a work alias, an English name or an online ID. Where public professional information lines up, you can match them to the same person. Where it doesn’t, flag it as unconfirmed. Holding these lines is a mark of respect for the candidate, and it protects your own reputation too.