In 2017 I gave a glowing reference for a man I knew would fail in that job, and he failed in it, and a company spent about two hundred thousand dollars finding that out.

I'm fifty-two. I've been in operations management for twenty-six years, mostly mid-size manufacturing. The man — I'll call him D — worked under me from 2012 to 2016.

Here's what I knew about him, precisely.

D was good at about sixty percent of his job and catastrophic at one specific part of it, which was anything involving a deadline he had not personally set. Not lazy. Not dishonest.

He had a thing — I think, though I'm not a doctor and I'm not going to diagnose a man on the internet — where a date someone else assigned to him did not feel real to him until about four days before it, and then it became an emergency and three other people absorbed it.

I managed around it for four years. I did it by giving him the kind of work that has no hard external dates, and in that work he was genuinely excellent — he redesigned a process in 2014 that is still in use.

In 2016 he left for personal reasons, moved two states, and in 2017 a company called me for a reference on a role that was, specifically and entirely, a deadline role. Program coordination. External customers.

And I said he was great.

I want to be exact about what I said, because I didn't lie. That's the part that has kept this alive in me for eight years. Not one sentence I said was false. I said he was the most capable process thinker I'd worked with, which is true. I said he was reliable, which was true in the context I'd built for him.

I said I'd hire him again, which is also true — I would, into the job I had him in.

The recruiter asked one open question: "Any areas for development?"

And I said: "He's strongest when he owns the plan."

That's it. That is the entire sum of my warning. "He's strongest when he owns the plan." Nine words that mean, to anybody who already knows, exactly what the problem is, and that mean nothing at all to a recruiter who has eleven calls that day.

He got the job. He was gone in fourteen months. I know the shape of it because a former colleague of mine ended up there in 2019 and told me, not knowing I'd given the reference: missed launches, two accounts lost, and a team that fell apart.

And here's what I actually did with it.

I did it because he had two kids and a mortgage in a new state and he had been looking for nine months, and because when he asked me to be a reference he said one sentence on the phone — "I need this one" — and because I liked him, and because telling the truth would have been me, in Ohio, with a job, ending something for a man in Tennessee without one.

I made a trade and I made it in about four seconds, the way you make all of these.

What I believe now, eight years on. The cost didn't land on the company, not really — a company that hires on one reference call deserves what it gets, and two hundred thousand is a rounding error to them. The cost landed on D. He had fourteen months of failing in public, and a résumé with a fourteen-month hole in it at forty-four, and whatever that did to how he thinks about himself.

And he would have been fine in a different job that I could have steered him toward if I had been honest for ninety seconds.

I protected him from a bad afternoon and I handed him a bad year.

I've changed exactly one thing. When I give a reference now — and I give maybe four a year — I say at the start: "I'll tell you what he's great at and where he'll struggle, and you should want both." And then I do it, by example, concretely, with the thing the person is actually bad at.

I've had two people call me afterward, upset. Both of them got the job anyway.

And I have never called D. He's on LinkedIn. He's at a company now in a role that looks, from the title, like the kind he's good at. I hope that's true.