Loyalty Expires the Day You Change Bosses

Last Wednesday, a little past eleven at night, I was on the subway when a colleague pinged me in a group chat: a proposal needed to go up the next morning, could I take a quick look at the architecture for anything broken. I crouched by the subway exit and sent back a voice message pointing out a concurrency case that wasn't handled. Nobody mentioned it the next day. On the third day, that colleague pulled me aside and said the flag saved them from a production incident review.

It reminded me of who I used to be a few years back. Back then I thought the way to hold a position at a company was to work the longest hours, answer every call, never push back on anything. I called that loyalty, and figured it was a chip I could keep spending. Looking back now, loyalty has a shelf life that's shorter than I expected.

It only holds value under one condition: you're still at that company, in that seat. Change the environment and loyalty resets to zero — nobody gives you credit at a new company for having been obedient at the old one. What actually travels with you, what genuinely makes people come looking for you, is that sliver of judgment — seeing how something should be handled a beat before everyone else does.

This isn't something I worked out on my own. Social psychologists John French and Bertram Raven published a study in 1959 that split "power" into five sources: coercive, reward, legitimate, expert, and referent. The first three are all tied to a position — the ability to punish, to hand out rewards, to give orders to subordinates. All of it is granted by the company, and when the company takes the position back, that power vanishes with it.

Expert power is different. It lives in you, and it walks out the door with you the moment you're let go. Loyalty, at bottom, is a one-sided investment made under legitimate and reward power: you keep giving to the organization, hoping it remembers what you put in — but an organization doesn't keep a ledger of feelings. It keeps a ledger of whether it can use you right now.

What actually helped that night wasn't the posture of "answering at any hour." It was whether the sentence I sent was useful. If I'd dragged myself up at midnight and flagged the wrong problem, nobody would have remembered the effort.

There's a place this is easy to get wrong: working long hours and answering every call aren't meaningless — they just don't produce expert power on their own. They only create the chance for it to be seen. What's actually worth something is the sentence itself — "there's a concurrency case that isn't handled." Being online at eleven p.m. is only what gives that sentence a chance to be noticed.

And there's a real cost here — building expert power is much slower than building loyalty. Loyalty just requires being willing to spend the time. Expert power requires actually thinking one step further than everyone else, verifying it, and then being seen doing it at the right moment. Plenty of engineers are willing to answer at any hour. Far fewer are willing to work through a risk nobody asked them to, before anyone asks.

I still haven't fully broken the habit. My first instinct that night was still "I can't be the one who drops the ball on this" — attitude points came to mind first, and it was only afterward that I realized the judgment call was what actually carried the weight. The two habits are still fighting it out. That night, the one that won was mostly luck.