We analyzed which types of job descriptions were the most effective at generating highly qualified candidates for clients. Here are our findings over the course of 5,000 applications and 25,000 people viewing a job description.
1. Job descriptions written by hiring managers or copywriters working closely with hiring managers were the most effective. In tests on Indeed, these postings generated more than twice as many highly qualified candidates as job descriptions for the same positions written by HR. Firsthand experience with the position and clear, persuasive writing are really helpful.
2. Shorter job descriptions are vastly more effective. In our tests, job descriptions 100-200 words long generated about 2-3 times as many qualified applicants as job descriptions longer than 500 words.
3. “Where should I start?” Our most effective template so far is a 1-2 sentence overview of the role and company, 3-5 main responsibilities, 3-5 important requirements, and 5-7 main benefits.
4. Picking the right requirements: focus on requirements necessary to succeed in the job, not “nice to haves.”
- Do most of your top performers in this position fit all of your stated requirements? If not, we’d suggest reevaluating what you think you need. (If your top performers generally don’t have X, requiring X is probably counterintuitive).
- We only recommend requiring a particular degree or major if it’s critical to job performance.
- Our candidate quality was generally higher on descriptions with fewer requirements.
5. Picking the right responsibilities: focus on critical, everyday tasks. You can discuss secondary responsibilities in more detail over the phone or during the interview.
6. Respond promptly to highly qualified candidates. They’re good enough to get hired quickly elsewhere.
7. Spend 5-10 seconds reading each application rather than relying on software screening. First, screening software incorrectly eliminates more qualified candidates, including some elite candidates). Second, the labor time to have a human scan for relevant job titles and/or experience is negligible. E.g. if your position receives 1,000 applicants, spending 5-10 seconds on each would only take 1-2 hours. If you do use screening software, I’d recommend testing that it correctly identifies resumes from employees currently in the position (especially top performers) as promising. Screening software can be extremely sensitive to phrasing, such as potentially eliminating extremely-qualified candidates if they use the phrase “Job History” rather than “Work Experience.”