What Building a Job Tool Taught Me About Hiring

2Most rejections are decided before a human reads anything

The common advice is that your resume needs better keywords. That is not wrong, but it describes the symptom. What actually happens is that a posting encodes a set of requirements, and your history either satisfies enough of them to clear a threshold or it does not. The keyword advice is a crude proxy for that.

When you score the same resume against a hundred postings, the pattern is obvious. The same person is a strong candidate for some roles and a weak one for others, and it has almost nothing to do with how the resume is worded. It has to do with which requirements are load-bearing for that specific role.

• Depth beats breadth: three years of one thing usually outscores one year each of three things.
• Recency matters more than people expect: a skill you used five years ago carries less weight than the same skill last year.
• Some gaps are fatal and some are noise: missing the core requirement sinks you, missing a nice-to-have rarely does.

3The listing itself tells you more than the company intends

A job posting is a document written under pressure by someone who has other work to do. That makes it leaky. A posting that lists fourteen required skills is usually a team that has not agreed on the role. A posting that has been reposted three times is telling you something about what happened to the last three candidates, or to the budget.

I ended up building a separate tool just for that signal, because it kept coming up: people wanted to know what it meant when a job they interviewed for showed up again a week later. The honest answer is that it depends on the timing and the wording, which is exactly the kind of thing worth checking rather than guessing about.

4Scoring changed what I applied to, not how I wrote

I expected the tool to make me a better applicant. What it actually did was make me a pickier one. Once you can see that a role scores in the fifties for you, the rational move is not to polish the cover letter. It is to spend that hour on the role that scores in the eighties.

That is an uncomfortable conclusion if you have been taught that effort is the lever. Volume and polish are the levers people reach for because they feel controllable. Selection is the lever that actually moves the outcome, and it requires admitting that some applications were never going to work.

5What I would tell someone starting a search now

Read the posting like a document with an author and a deadline, not like a specification handed down from somewhere. Ask which requirement is the one they cannot flex on. If you cannot tell, that uncertainty is itself information about how well defined the role is.

Then be honest about your own gaps before someone else is. Knowing that you are short on one specific thing is far more useful than a general feeling that you are underqualified, because the specific version tells you what to do next.

• Apply to fewer roles with more attention rather than many with less.
• Track which applications actually convert, not how many you sent.
• Treat a vague posting as a yellow flag, not a puzzle to solve on your own time.

6Why I am writing this here

ShouldApply came out of my own job search, and it is one of several things I have built over the years across pretty different industries. If you are curious about the rest of that work, it lives at digitaljesse.com.

I mention it because the useful thing I took from building in public is not a growth tactic. It is that shipping something and watching real people use it will correct your assumptions faster than any amount of reading about the problem. That applies to job searching too.