Writing Job Descriptions: How to Define Criteria That Actually Help

The average job posting receives 250 applications. Screening them takes up to 18 hours. Most of that time is wasted — because the job description was vague from the start.

Poorly written criteria create a domino effect. Recruiters spend hours reading CVs that never had a chance. Hiring managers reject 80% of the longlist. Candidates apply for roles they don't fit. Everyone loses.

Here's how to write job descriptions with criteria that actually work.

Why Criteria Quality Determines Screening Quality

A job description is not a wish list. It is a scoring rubric. Every criterion you include tells your screening process — human or AI — what to look for and how to rank candidates.

When criteria are vague, screening becomes subjective. One recruiter reads "strong communication skills" and looks for presentation experience. Another looks for published articles. A third skips it entirely because it's unmeasurable.

AI screening tools like HireSift face the same problem. They match CVs against criteria. If the criteria are fuzzy, the scores are fuzzy too.

The fix is simple: make every criterion measurable.

The 5 Most Common Mistakes in Job Descriptions

1. Using Adjectives Instead of Numbers

"Extensive experience in project management" means nothing. 3 years? 10 years? Does managing a team of 2 count?

Write instead: "5+ years of project management experience with teams of 10+."

2. Listing 20+ Requirements

Job descriptions with more than 12 criteria discourage qualified candidates. Research from LinkedIn shows that women apply only when they meet 100% of requirements. Men apply at 60%.

Keep your list to 6-8 criteria. Split them into must-haves and nice-to-haves.

3. Mixing Hard Skills and Personality Traits

"Proficient in SAP, team player, detail-oriented, SQL, proactive." This list mixes measurable skills with unmeasurable traits. It confuses both humans and algorithms.

Separate technical requirements from cultural expectations. Screen for skills first. Assess personality in interviews.

4. Copy-Pasting from Other Postings

Generic descriptions attract generic applications. If your posting reads like every other one on StepStone, expect the same flood of unqualified CVs.

Tailor criteria to the actual role. Talk to the hiring manager. Ask: "What does this person do in week one?"

5. Forgetting to Prioritize

Not all criteria are equal. A Java developer who lacks Kubernetes experience can learn it in 3 months. A candidate without Java cannot.

Weight your criteria. This is where AI scoring becomes powerful.

How to Write Measurable Criteria

Every good criterion answers three questions:

1. What is the skill or qualification?
2. How much is required (quantity, duration, level)?
3. How important is it relative to other criteria?

Here are examples across different criterion types:

Numeric Criteria

Vague	Measurable
"Experience in sales"	"3+ years B2B sales experience"
"Knowledge of accounting"	"2+ years working with IFRS standards"
"Team leadership"	"Led teams of 5-15 people"

Boolean Criteria (Yes/No)

• "Valid German nursing license (required)"
• "Willing to relocate to Munich (required)"
• "EU work permit (required)"

Enum Criteria (Specific Levels)

• "German language: C1 or higher"
• "Education: Master's degree in Computer Science or equivalent"
• "Certification: PMP or PRINCE2"

Multi-Select Criteria

• "Tech stack: At least 3 of the following: Python, Java, Go, Rust, TypeScript"
• "Industry experience: Healthcare, Pharma, or MedTech"

Must-Have vs. Nice-to-Have: The 60/40 Rule

Split your criteria into two tiers. Aim for 60% must-haves and 40% nice-to-haves.

Must-haves are non-negotiable. A candidate without them cannot do the job on day one. Examples: required certifications, minimum experience levels, language requirements.

Nice-to-haves are differentiators. They separate good candidates from great ones. Examples: specific industry experience, additional tools, leadership experience.

This split serves two purposes. First, it widens your applicant pool. Candidates who meet 60% of criteria will apply. Second, it enables nuanced scoring. AI tools can weight must-haves at 2-3x the value of nice-to-haves.

How Criteria Feed Into AI Scoring

Modern AI screening tools use your criteria as scoring dimensions. Here's how HireSift processes them:

1. CV Match Score: Measures how well a candidate's CV matches your stated criteria. Each criterion gets a weight. The algorithm checks for matches across experience, education, skills, and certifications.

2. HireSift Score: Goes beyond simple matching. It evaluates overall candidate quality, career progression, and relevance — things a good recruiter would notice but that take time to assess manually.

Both scores depend entirely on your criteria quality. Better criteria produce more accurate scores. Vague criteria produce noise.

A Template for Better Job Descriptions

Use this structure for every role:

Role: [Title]
Department: [Team]
Reports to: [Manager title]

Must-Have Criteria (weighted 70%):
1. [Skill] — [Measurable requirement] — Weight: [High/Medium]
2. [Qualification] — [Specific level] — Weight: [High/Medium]
3. [Experience] — [Years + context] — Weight: [High/Medium]

Nice-to-Have Criteria (weighted 30%):
1. [Skill] — [Measurable requirement] — Weight: [Medium/Low]
2. [Experience] — [Years + context] — Weight: [Medium/Low]

Example: Senior Backend Developer

Must-Have Criteria:
1. Java — 5+ years production experience — Weight: High
2. Microservices — Built and maintained 3+ services — Weight: High
3. SQL databases — PostgreSQL or MySQL, 3+ years — Weight: Medium
4. CI/CD — Experience with Jenkins, GitLab CI, or GitHub Actions — Weight: Medium

Nice-to-Have Criteria:
1. Kubernetes — 1+ year in production — Weight: Medium
2. Event-driven architecture — Kafka or RabbitMQ — Weight: Low
3. Mentoring — Led or mentored junior developers — Weight: Low

The Bottom Line

Writing job descriptions is not an administrative task. It is the single most important step in your recruiting process. Every minute spent on clear, measurable criteria saves 10 minutes in screening.

Define what matters. Quantify it. Weight it. Then let your screening process — whether human, AI, or both — do what it does best: find the right people, fast.

For more on how AI scoring works in practice, read our guide on AI recruiting in the tech industry. And if you're preparing for the next step after screening, check out our article on interview preparation.

---

Less screening. More hiring.

HireSift analyzes 100 CVs in minutes — with two transparent scores, EU AI Act compliant, no credit card required.

→ Start your 7-day free trial