Writing Inclusive Job Descriptions That Actually Work

Every open role has an invisible filter that most hiring teams never audit: the job description itself. Before a single resume lands in your ATS, the language you chose — the requirements you listed, the culture you described (or didn't), the adjectives you reached for — has already determined who applies and who closes the tab.

The data is striking. A 2024 study published in PNAS found that replacing masculine-coded language in job advertisements with gender-neutral synonyms meaningfully increased the diversity of applicant pools. Earlier research showed that postings heavy with male-coded words like "dominant," "aggressive," and "competitive" saw applications from women drop by up to 10%. And a large-scale analysis of 60,000 job descriptions found that companies that rewrote their postings for inclusive language saw total applications increase by 13% and applications from women rise by over 20%.

The problem is not just gendered language. It extends to age-coded phrases ("digital native"), ability assumptions ("must be able to lift 50 lbs" for a desk job), inflated requirements lists, and culture descriptions that signal exclusion while intending to signal fun. And the cost is not abstract — it's measurable in a smaller, less diverse, and ultimately less qualified candidate pool.

We analyzed job descriptions across 118 companies in our Culture Directory and cross-referenced them with employee review data and application patterns. This guide distills what actually works — not the surface-level advice you've read before, but specific, research-backed changes that move the needle.

The Language Problem

Most exclusionary language in job descriptions is not intentional. No hiring manager sits down thinking, "Let me make sure women don't apply." The problem is that certain words and phrases carry gendered, aged, or ability-based connotations that researchers call "linguistic cues" — signals that tell candidates whether they belong before they've read the actual job requirements.

Masculine-coded words like "rockstar," "ninja," "dominate," "crush it," and "aggressive" are the most studied examples. A foundational study by Gaucher, Friesen, and Kay (2011) demonstrated that job advertisements with more masculine wording were perceived as less appealing by women, who also felt a lower sense of belonging in the described workplace. More recent work published in PNAS (2024) confirmed that debiasing these terms increases gender diversity in applicant pools without reducing overall application volume — you don't lose male applicants by dropping "ninja."

But gendered language is only part of the picture. Here are the most common categories of exclusionary phrasing we found across real job postings:

The pattern is consistent: replace vague, culturally-loaded language with specific, outcome-oriented descriptions. "Rockstar" tells a candidate nothing about what they'll actually do. "Senior Software Engineer who will own our payments infrastructure" tells them everything. Specificity is inherently more inclusive because it lets candidates evaluate fit based on skills and experience rather than identity signals.

Requirements vs. Nice-to-Have

You've probably heard the widely-cited claim that men apply for jobs when they meet 60% of the qualifications while women only apply when they meet 100%. The original statistic traces back to anecdotal comments from a Hewlett Packard internal report, not rigorous research. But when researchers at the Behavioural Insights Team and later in a 2024 European Journal of Social Psychology study tested this claim rigorously, they found the underlying pattern is real — just smaller than the viral version suggests.

Here is what the research actually shows: women are somewhat more likely than men to self-select out of applying when they don't meet every listed requirement. The gap narrows significantly when requirements are clearly separated into "must-have" and "nice-to-have" categories. In other words, the structure of your requirements list matters as much as the content.

Our analysis across companies in the JBC directory found a clear pattern: the best-performing job descriptions list 5–7 core requirements and explicitly label everything else as preferred or bonus qualifications. Here's how to restructure yours:

• Hard requirements = skills the person literally cannot do the job without. If your codebase is in Go, "proficiency in Go" is a requirement. If you use Kubernetes in production, "Kubernetes experience" is a requirement. The test: could someone learn this in the first 90 days? If yes, it's not a hard requirement.
• Nice-to-haves = everything else. Experience with a specific CI/CD tool, familiarity with a particular database, exposure to a certain domain — these are preferences, not gates. Label them explicitly: "Bonus: experience with Terraform" or "Nice to have: familiarity with healthcare data."
• Cap hard requirements at 5–7. Every requirement beyond that narrows your pool. If you genuinely need 15 specific skills, you either need two roles or you haven't thought clearly about what's truly essential for day one versus day ninety.
• Replace years-of-experience gates with demonstrated competency. "7+ years of Python" tells you nothing about capability. "Strong experience building and maintaining production Python services" tells you what matters. A developer with 3 years of deep, ownership-level Python experience is often more capable than one with 10 years of surface-level scripting.

The companies with the most diverse engineering teams in our directory — including HubSpot, Spotify, and Notion — all follow this pattern. Short, honest requirements lists. Clear separation of must-haves from nice-to-haves. No year-count gates. No keyword-stuffing.

Show Culture, Don't Just List It

The weakest section of most job descriptions is the "About Us" paragraph. It usually reads like a press release: "We're a fast-growing startup disrupting the [X] industry with a passionate team of [Y] people." This tells candidates almost nothing about what it's actually like to work there — which is exactly the information that drives the apply-or-pass decision for the best candidates.

We built the JBC culture values system to solve this problem at scale. Every company in our directory is tagged with verified culture values — Remote-Friendly, Work-Life Balance, Engineering-Driven, Flat Hierarchy, Async Culture, Learning & Growth — based on employee reviews, concrete policies, and observable evidence. Not marketing copy. Not aspirational values posters. Documented reality.

Your job description should do the same thing. Here's how to translate abstract culture claims into specific, believable signals:

• "We value work-life balance" is meaningless. "Our engineering team averages 42-hour weeks. On-call rotations are one week per quarter with compensatory time off. We haven't had an all-hands after 5pm in two years" — that's a culture signal a candidate can evaluate.
• "We're remote-friendly" could mean anything. "70% of our engineering team works remotely. We default to async communication via Notion docs and Loom videos. Meetings are batched to Tuesday/Thursday afternoons" — that's verifiable.
• "Great engineering culture" is a claim. "Engineers write design docs before code, ship to production 12+ times per day via CI/CD, and choose their own tools within our infrastructure guardrails" — that's evidence.

The principle is simple: replace adjectives with specifics. Instead of saying your culture is "collaborative," describe how collaboration actually happens. Instead of calling your team "diverse," share your actual demographic data or describe your ERG programs. Candidates — especially those from underrepresented groups — have learned to distrust adjectives. They trust mechanics.

Benefits That Signal Inclusion

Benefits are not just perks — they are the strongest signal of who your company is actually designed for. A generous parental leave policy tells candidates with families (or planning families) that they won't be penalized for having children. Flexible hours tell candidates with disabilities, chronic conditions, or caregiving responsibilities that they can structure work around their needs. Remote options tell candidates outside your city's demographic bubble that they're welcome.

The companies that attract the most diverse candidate pools in our directory share a specific set of benefit signals. Here's what to highlight — and how:

• Parental leave: State the specific number of weeks and whether it's gender-neutral (not "maternity leave"). Companies like Spotify offer 6 months of paid parental leave regardless of gender — and they put it in job descriptions because it's a competitive advantage.
• Flexible/async hours: Specify whether "flexible" means "you choose your 8 hours" or "core hours 10-3 with flexibility around those." Vagueness here breeds distrust. Companies tagged with Flex Hours in our directory are explicit about what flexibility actually looks like day-to-day.
• Remote work: State clearly whether the role is fully remote, hybrid, or on-site. If hybrid, state which days are required. If remote, state any timezone constraints. Browse companies tagged Remote-Friendly in our directory for examples of how to do this well.
• ERGs and communities: If you have Employee Resource Groups, name them. "We have ERGs for women in engineering, LGBTQ+ employees, and parents" is a concrete signal. "We value diversity" without specifics is noise.
• Professional development: State the specific budget. "$3,000 annual learning stipend for conferences, courses, or certifications" is more believable than "we invest in your growth." HubSpot is particularly strong here, with transparent career frameworks and dedicated L&D budgets.
• Mental health and wellness: Therapy stipends, mental health days, wellness programs — if you offer them, list them. These benefits disproportionately matter to candidates who have been burned by "hustle culture" workplaces.

The pattern across the companies that do this best: be specific, be honest, and lead with the benefits that matter most to underrepresented candidates. Those same benefits, incidentally, are the ones that every employee values most — inclusive benefits are better benefits.

Before/After Examples

Theory is useful. Examples are better. Here are three real-world rewrites that show how small changes in structure and language can transform a job description from exclusionary to inclusive — without sacrificing clarity or standards.

Notice the patterns across all three rewrites: specific titles instead of superlatives, outcomes instead of adjectives, separated requirements, explicit comp ranges, and honest descriptions of the work arrangement. Every change makes the posting more informative for everyone while removing signals that disproportionately deter candidates from underrepresented groups.

Tools for Auditing Your JDs

Even with the best intentions, unconscious bias in language is hard to catch manually. Fortunately, several tools exist to help hiring teams systematically audit their job descriptions for exclusionary language. Here are the most effective ones we've seen companies in our directory use:

• Textio — the most sophisticated option. Uses AI and a proprietary dataset of hiring outcomes to score your job descriptions for bias across gender, age, and ability dimensions. It doesn't just flag problems; it suggests specific replacements and predicts how changes will affect your applicant pool composition. Used by many enterprise teams across our company directory.
• Gender Decoder — a free, lightweight tool based on the Gaucher, Friesen, and Kay (2011) research. Paste your JD and it identifies masculine and feminine-coded words. It's less sophisticated than Textio but it's free, fast, and catches the most obvious patterns. A good starting point for teams without a budget for premium tooling.
• Ongig Text Analyzer — scans for exclusionary language across multiple dimensions: gender, age, ability, race, and mental health. Also checks reading complexity and suggests simplifications. Particularly useful for companies writing at scale (50+ open roles at a time).
• Hemingway Editor — not specifically designed for bias, but useful for readability. Job descriptions written at a high reading level exclude candidates who are non-native English speakers or who have different educational backgrounds. Aim for grade 8–10 reading level.

However, tools have limits. They catch keyword-level bias ("ninja," "aggressive") but miss structural bias (a requirements list that's 15 bullets long) and contextual bias (describing your office culture in a way that assumes everyone is 25 and childless). The best approach combines automated scanning with review from diverse team members who can spot assumptions that algorithms miss.

For a deeper look at how the best companies structure their job descriptions, see our guide on writing engineering job descriptions that senior engineers actually read.

Making It Stick: An Auditing Checklist

Writing one inclusive job description is a start. Building a system that produces inclusive descriptions by default is the goal. Here's a practical checklist to run against every JD before it goes live:

1. Language scan: Run the description through a tool (Gender Decoder at minimum, Textio if available). Flag and replace any masculine-coded, age-coded, or ability-coded language.
2. Requirements audit: Count your hard requirements. If there are more than 7, cut. For each remaining requirement, ask: "Could someone learn this in 90 days?" If yes, move it to nice-to-have.
3. Year-count purge: Replace every "X+ years of experience" with a competency description. "5+ years of Python" becomes "strong production Python experience."
4. Culture specifics: Check that every culture claim is backed by a specific, verifiable detail. No adjectives without evidence.
5. Benefits prominence: Ensure parental leave, flexibility, remote policy, and comp range are visible — not buried at the bottom.
6. Pronoun check: Replace all instances of "he/she" with "you" or "they." Better yet, address the candidate directly throughout: "You will..." instead of "The candidate will..."
7. Diverse review: Have at least one person who is not the hiring manager read the JD before posting. Ideally, include a reviewer from an underrepresented group.

If you're an employer looking to attract candidates who care about culture, the first step is making sure your job descriptions reflect the culture you've actually built. Companies that list on JobsByCulture go through a culture verification process that ensures what you say matches what employees experience. Learn how to list your company and reach candidates who filter by values, not just keywords.