Harvard Business Review published research last week that arrives looking like somebody else’s problem. The researchers analysed fifty million European job postings and ran experiments with twelve hundred hiring managers. A remote role, they found, now asks for roughly twenty five percent more skill, experience and credentials than the same role performed in person, and firms are increasingly hiring only people who arrive needing no development at all. Note what was measured. Not whether remote work is productive. What an employer now requires before it will consider you at a distance. And underneath the coverage sits a belief the finding appears to leave alone, the one a senior professional holds with some confidence: my setup works. Fewer interruptions, better hours, more done. The best working arrangement I have ever had, and my own output is the evidence.
The first level read splits the usual two ways. One camp treats it as a warning about entry level opportunity: employers should recalibrate their criteria, rebuild the mentoring that used to happen by itself, stop asking for five years of experience for a job that teaches itself in one. The other camp waves it off: European postings, a labour market still settling after five disrupted years, and the junior end always takes the first squeeze. Notice what the two camps share. Both read it as a hiring problem, decided inside somebody’s recruitment function, happening to people who are not in the room. Neither asks what it costs the man whose setup is working.
It is not a hiring problem. It is an accounting problem, and the research points straight at the ledger. Development used to be paid ambiently, by everybody at once and nobody in particular. The overheard call. The correction that took forty seconds because you happened to be standing there when it went wrong. The meeting a junior sat in on for no stated reason and left knowing how the decision had actually been made, which is the part nobody writes down. None of it was scheduled, none of it appeared on anybody’s day, and nobody could have told you what it cost, because nobody was carrying it separately. Remote did not destroy that. Remote itemised it. It turned something invisible and shared into a line item somebody would now have to own, schedule and defend. And a line item nobody agreed to own gets routed around.
Both routes are rational, and neither involves anybody behaving badly, which is why this is worth understanding rather than arguing about. The firm routes around it by asking for people who arrive already made, and that is defensible on any single hire. The senior professional routes around it by protecting the day that finally works, and that is defensible on any single afternoon. His focus is real and it is valuable. The deep work genuinely is better than it was across a desk from three other conversations. Nobody ever told him that developing people was part of what he was paid for, because when it was done to him it was ambient too, so he assumes it is still happening somewhere. And the routing happens one correct decision at a time. Something comes back slightly wrong, and where you would once have said a single sentence across a desk, saying it now means finding a slot and turning forty seconds into twenty minutes, so you fix it yourself in four. Twenty of those, and there is a person who has not learned the thing.
I have watched the receiving end of this many times. On a trading floor a man’s decisions are fed by things nobody scheduled: the conversation two rows back that he is not part of and is half hearing, the way the room moves when a number prints, the fact that someone he respects has gone quiet. A fair amount of it is noise and I would not pretend otherwise. When we moved traders to working remotely we did it slowly and never as a cliff edge, and the reason was always the same one. The trader does not know what he is about to lose. He knows he will miss the company. He does not know how much of that floor had been running inside his own decisions, because none of it ever arrived labelled as an input, and the work after that is learning to decide as well without it. We have put real thought into rebuilding that room digitally and we have not managed it. Everything we have tried reproduces the talking and not the room.
The durable truth under the research has nothing to do with where anybody sits. A cost that was always being paid invisibly does not stop being paid when you stop being able to see it. It gets reassigned, and it goes to whoever is least able to refuse it: to the people who never got into the room, and, quietly, back to the man who left it. Its partner is a rule about scoreboards. What a senior person is worth is not what he produces, it is the difference he makes to what everyone around him produces, and that second figure appears on nothing he looks at. Which is the trap in a sentence. It is possible to optimise your way back into being someone who only does his own work, and to do it while your own numbers are the best they have ever been.
So the question the research leaves you with is not whether remote work is good, which is a policy question and belongs to somebody else. It is narrower than that, and it is yours. Of the work you did last week, how much of it raised what somebody else is capable of, and how much of it was simply your own, done well? And if that balance has been drifting one way for two years, which of the numbers you actually look at would have told you?