FROM THE DESK
Twenty-two years of placing people, written down. Briefs, screening, candidate care, and the working detail that decides whether a placement sticks.
Section 71 prohibits decisions based solely on automated processing that profile performance at work. Most recruiters have never read it.
The advice to delete every unsuccessful applicant is wrong, and section 14 says so. What you may actually keep, and on what basis.
Most briefs describe a job that does not exist. The three questions a specification never asks.
Keyword matching returns candidates who are plausible and unsuitable, with a percentage attached. Amazon spent four years learning this.
A transcript records words. A recruiter records meaning, and the difference decides placements.
They did not lose interest. They lost the thread. Four points where good people quietly disengage.
Automation multiplies whatever it is pointed at. A committee wish list produces a wish list at scale.
Structured interviews predict performance roughly twice as well. Structure is what creates room for the conversation.
Ranking optimises for the brief. People optimise for the team, the manager and the next two years.
Three CVs that miss is not a wasted week. It is a client quietly briefing someone else.
The most valuable thing a recruiter says is often that the search should stop. In week one it is judgement; in week six it is an excuse.
The statistic every recruiter quotes has no traceable source. The argument is stronger without it.
They predict performance at about 0.23. That is a reason to run them differently, not to skip them.
What the BCEA actually requires, and why the contract usually says something longer.
A model that invents a job title has not made a quirky error. It has misrepresented a person to an employer.
A straight ledger of what to automate and what to protect, from a desk that automated its own work first.
A recruiter forms private judgements about people they are paid to represent. Where the line falls is a craft nobody writes down.
Everyone counts the replacement cost. Almost nobody counts the good person who left because of it.
Years of experience correlates with performance at about 0.18. The most-used filter is among the least useful.
Try a different word, or clear the filters to see all nineteen articles.
We are taking our first 20 customers, with hands-on setup.