Every recruiter has been handed the spec. Eight years of experience, three languages, a niche certification, startup pace with enterprise rigor, and a salary band set two years ago. Somewhere out there, the theory goes, one person matches all of it.
Recruiters call that person a purple squirrel. The name is, of course, a bit of fun, but the joke isn’t really about this mythical candidate. It’s about the ridiculous job spec. That distinction matters more in 2026 than it did when the term was coined, because the thing that used to make purple squirrels rare has stopped working.
This piece covers what the term actually means, why these job descriptions keep getting written, what chasing one costs you, and what to do instead.
A purple squirrel is a candidate who would match every requirement on a job description exactly, including the requirements that quietly contradict each other. The term describes the improbability of the combination, not the seniority of the person.
Recruiters have used it since the early 2000s, and they use it about their own hiring managers. It is shorthand for a request that cannot be filled as written. When a recruiter tells you your role is a purple squirrel hunt, they are not complimenting your standards. They are telling you the document is broken.
Two things the term does not mean, both of which get muddled regularly.
It does not mean a passive candidate. Whether someone is currently job hunting has nothing to do with it. You can find a passive candidate in a week. You cannot find a person who does not exist, however long you look.
It does not mean a high performer. Plenty of exceptional people fail a purple squirrel spec (say that five times fast), because the spec was never calibrated against anyone real. Failing it is evidence about the document, not about the candidate.
Rare skill combinations exist. People who satisfy every line of an uncalibrated spec, at the. The useful question is not whether the person is out there. It is whether your spec describes a job or a wish list. Those are different documents, and only one of them can be filled.
Most purple squirrel specs get assembled the same way. Someone lists everything the last three people in the role could do between them, adds whatever the team is currently missing, then sets a salary based on the last time the role was filled. Each addition is reasonable on its own. Together they describe a composite of several careers.
The market is not withholding this person from you. The document is asking for someone who was never going to turn up.
Biweekly updates. No spam. Unsubscribe any time.
They are rarely written in bad faith. Four things produce them, and recognizing which one you are dealing with tells you how to fix it.
The requirements describe what a team needs in total, not what one role delivers. This happens most often when a team has been under-resourced for a while and a single headcount finally gets approved. Every unmet need in the department attaches itself to that one requisition.
A hiring manager who cannot afford a mis-hire raises the bar past the point of realism, because a spec nobody meets feels safer than a judgment call they might have to defend. The requirements are not really requirements. They are insurance.
Nobody has agreed what good actually looks like in the role, so credentials get added as stand-ins. Years of experience. A degree. A former employer. A certification. Each proxy narrows your pool without improving the quality of the decision, and the effect compounds. Five proxies can remove almost everyone while telling you almost nothing.
The spec never gets tested against a real person before it goes live. Nobody finds out it is unfillable until month three, by which point the vacancy has its own momentum and the requirements have become load-bearing.
The cost is not only the open seat. It is what the search does to everything around it.
Every applicant is measured against an ideal rather than against the job. People who would have succeeded get filtered out for missing a line item nobody needed, and the person doing the filtering feels rigorous while doing it. This is the most expensive cost and the hardest to see, because rejected candidates do not send you performance data.
Stacking proxies removes candidates by background rather than by ability. That is how a search for excellence ends up producing a shortlist that looks a lot like the team you already have, and it is a well-documented route to bias in hiring even when nobody intends it. You did not decide to narrow the field on background. You decided it by proxy.
Existing employees cover the open role for as long as the search runs. A long time to hire is not a neutral delay. It is workload moved onto people who did not volunteer for it, which is its own retention risk.
If you do land someone that specialized, so can your competitor. The rarer the profile, the more portable it is. You spent six months on a hire whose market value you just publicly confirmed.
The purple squirrel used to be a search problem. It is now an evidence problem, and that is a different job.
Candidates write resumes with AI, and AI is very good at producing a document that matches a job description. It reads the posting, mirrors the language, and returns a profile that hits every requirement. The improbable candidate is no longer rare. On paper, the improbable candidate is the default output.
So the signal has inverted. A profile that matches your spec perfectly used to mean you had found someone remarkable. It now means someone read your posting carefully. We covered the mechanics of this in how AI-written resumes broke candidate sourcing.
This is why the traditional advice has stopped helping. Search harder, network wider, work more channels. All of it now produces more perfect-looking profiles and no more certainty about any of them. You cannot out-search a problem that lives in the evidence layer.
Five steps, in order. The first three take an hour with the hiring manager and save you the argument you would otherwise have in week six.
Put down what this person has to be able to do in their first ninety days. Not the title you would like them to have held. Not the years you would like them to have served. Capabilities can be tested. Credentials are proxies, and proxies are what broke.
Do this before you search, not after. Must-haves decide who makes the list. Nice-to-haves decide the order. Sorting them upfront means you stop renegotiating the shortlist later, when the hiring manager has a name in front of them and a preference they are about to call a requirement.
If everything is essential, nothing is. Most roles have two or three genuine deal-breakers. The rest is preference wearing a stricter word. Forcing the cap is uncomfortable, and it is the step that does the most work.
For each must-have, decide now what would settle it. A score on a relevant skills test. A job simulation. A work sample. Do this with the hiring manager in the room, because the disagreement you have here is the one that otherwise surfaces after three rounds of interviews.
Take someone already succeeding in the role, at your company or a comparable one, and run them against your requirements. If they fail, the spec is wrong. Not the market, not your sourcing, not the candidates. The document.
Once the spec describes capabilities, the search changes shape. You stop hunting for a profile that matches on paper and start looking for evidence that someone can do the work.
In practice, that means three things. Everyone you seriously consider gets assessed against the same standard, so the comparison actually holds. The hiring manager sees a score next to each name rather than a resume and a recommendation. And the shortlist can be defended, because every name on it is there for a stated, checkable reason.
That last point is what ends the purple squirrel conversation for good. The reason unfillable specs survive is that nobody can prove the alternative candidate would have worked out. Resume screening cannot settle that argument. Evidence can.
A work history suggests. Only an assessment shows.
CTA. Stop writing specs nobody can fill. Build an assessment around what the role actually needs, and see who can do it. Try TestGorilla for free or book a demo.
It describes a candidate who would match every requirement on a job description exactly. Recruiters use it about job specs that cannot be filled as written, not about candidates themselves.
It has circulated among recruiters since the early 2000s and spread during the 2008 downturn, when employers wanted single hires who could cover work previously done by several people.
The terms are used interchangeably. Both describe an unrealistic spec rather than a type of person.
Writing a demanding spec is legal. The risk comes from requirements that screen out protected groups without predicting performance, which is why capability-based requirements sit on firmer ground than proxy-based ones. Fair hiring practices covers this in more detail.
Define what the role delivers in the first ninety days, cap your must haves at two or three, agree what proof looks like for each, and test the finished spec against someone already doing the job well.
Why not try TestGorilla for free, and see what happens when you put skills first.