Fewer Applications, Better Ones: What Four Edits to a Job Advert Changed in Three Weeks

A five-person machine shop needed one more operator and had been running the same advertisement for eleven weeks. It produced applications steadily, roughly forty of them, of which two were worth a phone call and neither worked out. The owner's reading of that was that the labor market was empty, which is the reading almost everybody reaches, and it is usually the wrong one. The advertisement was doing something more specific than failing to attract people. It was attracting the wrong ones efficiently, and it was quietly repelling several of the right ones.

What the Rewrite Changed

Four things, none of them stylistic. The first was the title, which had read Manufacturing Associate and now read the name of the machine the person would actually be running. Nobody looking for that work searches for an internal job classification, and the previous title had been drawing applications from people who had no idea what the role involved because the words did not describe anything in particular.

The second was pay. The old notice said competitive, and the new one gave a range with a note about what moved somebody through it. The third was schedule, which had been unstated and which turned out to be the single largest filter, since the shop runs a four-day week with long days and that is either exactly what somebody wants or completely impossible for them. The fourth was the requirements list, which is where most of the damage had been done and which needs its own section.

What Happened to the Numbers

Applications fell by roughly two thirds over the next three weeks. Phone screens rose from two in eleven weeks to seven in three, in-person interviews from zero to four, and the shop hired from the second round. The owner's initial reaction to the falling application count was alarm, which is the standard reaction and the reason most adverts drift back toward vagueness over time. Volume feels like reach. In hiring it is mostly noise, and every unsuitable application costs somebody twenty minutes and a small amount of goodwill on both sides.

The more interesting number was the change in who applied. The old notice drew people with no relevant experience who were applying broadly, plus a few who were plainly overqualified and looking for anything. The new one drew people who had run similar machines, several of whom mentioned the schedule specifically in their first sentence. Publishing the two facts that decide whether somebody can take the job produced a pool that had already made that decision before applying.

The Requirements List, Which Was Doing Damage

The old advert listed eleven requirements, assembled over several years by adding whatever the last hire had turned out to lack. Among them were a high school diploma, a valid driver's license, five years of experience, the ability to lift a stated weight, and familiarity with three specific software packages, two of which the shop no longer used. Almost none of these were genuine requirements. They were preferences, historical accidents and defensive additions, and together they read as a wall.

Requirements lists filter in two directions and only one of them is intended. They screen out people who cannot do the job, which is the point, and they also screen out people who can do the job but who read a list of eleven items, count the two they lack, and decide not to apply. Research on that asymmetry is well established and the effect is not evenly distributed across applicants. The rewritten list had four items, all genuinely necessary, with a separate short line naming two things that would help but were not required.

The Local Piece

Two local realities shaped the rewrite as much as anything on the page. The shop sits eighteen minutes outside a town with almost no public transit, which makes commuting distance a real constraint that the advert had never acknowledged. Stating the location plainly, including the drive time from the nearest population center, cost a handful of applications from people who would have withdrawn anyway once they mapped it.

The other was the local wage picture for comparable work. Publishing a range only helps if the range is credible, and the way to find out is to look at what similar work pays in the same metropolitan area rather than nationally. Occupational wage data broken down by region and by specific occupation is published by the Bureau of Labor Statistics, and reading it before writing the range is what stopped the shop from advertising a number that would have looked unserious to exactly the people it was trying to reach.

What Did Not Change

The job itself did not change, the pay did not rise, and the shop did not start offering anything it had not offered before. That is worth saying plainly, because the usual explanation for a hiring problem is that the offer is not good enough, and this offer had always been reasonable for the work and the area. What had been wrong was the description of it, which is a considerably cheaper thing to fix and one that most small employers never revisit because the advert was written once, worked at the time, and has been copied forward ever since.

The forty applications in eleven weeks were never evidence of a healthy pipeline, and the fourteen in three weeks were never evidence of a shrinking one. They were measurements of how well a piece of writing described a job to the people who could do it. A shop that treats its advert as a document to be edited rather than a form to be reused gets a different pool from the same town, the same wage and the same week, which is a fairly ordinary conclusion that takes an afternoon to act on and eleven weeks to arrive at the hard way.