
Picture this. Two companies post for the same role — a mid-level compliance analyst. Same salary range. Similar responsibilities.
Both descriptions are accurate. Both companies are probably fine places to work. But they are not saying the same thing.
Candidates know it.
Most hiring managers think applicants read for requirements. They don’t — not primarily. They read for culture. A single adjective like “aggressive” or “dominant” triggers something closer to pattern recognition than conscious analysis.
This isn’t a soft observation.
Think about who that second group includes. Plenty of excellent analysts, engineers, and strategists who simply don’t perform well under the framing of “we want a warrior.”
You’re not filtering for competence. You’re filtering for self-presentation style. Those are very different things.
Let’s stay concrete, because this topic attracts a lot of vague claims.
Twenty-five percent faster. In a market where roles can sit open for 45, 60, or 90 days, that’s not a rounding error.
There’s also the Glassdoor dimension, which tends to get underweighted in these conversations.
You’ll never know these people existed. They just didn’t apply.
The honest synthesis: language functions as part of a coherent signal. When word choices, listed benefits, tone, and company description all tell a consistent story, candidates form a clear impression. Change just the adjectives while leaving everything else pointing in the old direction, and the effect is muted. That’s probably what MIT Sloan is observing.
Signal coherence is the real variable. Not any single word.
Not a checklist. More like a quick test you can run on any job ad before publishing: what kind of person would feel unwelcome here?
None of this means softening the actual requirements. It means precision. Describe what the role genuinely demands and what working there actually looks like. The candidates who self-select in will be better matches. The ones who self-select out? They would have struggled anyway.
Employer brand isn’t something you build on a careers page you redesign every few years. It’s built by every piece of communication a candidate reads before they ever speak to a recruiter.
The job posting is that communication. A posting written carelessly isn’t just neutral. It’s information. And candidates are good at reading it.
The issue is rarely malice. It’s inertia.
Most job postings are written by copying last quarter’s version and updating the headcount. The language accumulates — phrases that felt standard five years ago, requirements nobody questioned because whoever wrote them also wrote the last ones. A tone that persists not because anyone chose it, but because no one changed it.
Language risk in hiring documentation is often invisible until it shows up as a pattern: in your applicant data, in your Glassdoor ratings, in a candidate who quietly tells a recruiter the posting didn’t feel right for someone like them.
The good news is that fixing this is mostly an editing problem, not a culture problem. Start with the language. The signal changes fast.