We advertised for an infrastructure engineer three times over five months. Nine applications, two worth interviewing, one offer made and declined for eight thousand pounds more elsewhere. The band was set in 2023 and the benchmark that governs it is refreshed every two years, so the answer from human resources was correct, consistent and entirely useless to me.
The work still had to be done. So we engaged a contractor at five hundred and fifty pounds a day, which is roughly a hundred and thirty thousand a year, and he stayed fourteen months. Nobody queried it once. Not the committee, not finance, not the audit of professional services spend that happened halfway through. It was project expenditure, and project expenditure is governed by a budget. A permanent post is governed by a headcount ceiling, and the two are controlled by different people for different reasons that have never been reconciled.
We were not short of money for this work. We were short of a way to spend it on a person who stays.
The second cost took longer to see. He was good, he did the job, and he took the job with him. Our storage platform, our backup design and two of our automation pipelines were understood properly by one man with a notice period of two weeks. The handover was three weeks of screen sharing and a wiki that nobody else has edited since. Meanwhile the vacancy stayed open, the overtime bill stayed high, and one of my permanent engineers left in part because the gap had been his to carry for a year.
I present these decisions differently now. Every proposal for a role shows both routes costed over three years, fully loaded, including the cost of the vacancy itself, which we estimate rather than leave blank. Two contractors have been converted to employees on that basis. Three scarce skills carry a retention supplement argued as avoided cost rather than as a pay case. And any engagement running past six months must name an employee working alongside, with handover treated as a deliverable and held against the final invoice.
A hiring problem that dissolves the moment you rename the money was never a market problem. It was a governance shape.
– Serguey Shinder
