Great Companies Are Built on Talent Leaders Can See Accurately, Not Just Easily

Great companies are built on talent leaders can see accurately, not just easily.

Every company says it promotes on merit.

Then the promotion meeting starts, and someone says, "I just think she's ready."

Nobody opens a document, because there isn't one. There's only what people remember, and people remember what stood out.

That's salience bias. Picture it. The person presenting to the executive committee gets remembered. The person whose model quietly caught the pricing error doesn't, because nothing happened. The engineer who fixed the 2 a.m. outage gets thanked by name. The one who spent a year preventing outages gets nothing.

It isn't a small bug. In two large datasets of manager ratings, the rater's own quirks explained more than half the variance in the scores (Scullen, Mount and Goff). In a randomized experiment, remote workers who outperformed were promoted less (Bloom et al.). You'd never approve a budget because someone remembered a good slide. You'll promote on it.

You can't remove the bias. You can stop letting it make the decision. Here's where to start.

If you lead. Before the next promotion conversation, write down two things each candidate produced that have nothing to do with how well they present. If you can't, you don't know them yet. Then ask the question nobody asks: whose failure did they prevent?

If you run the company. No evidence file, no vote. Dated outcomes for the whole year, circulated 48 hours before the meeting. Three independent scores before anyone tells a story. Twelve months later, check whether the people you promoted were any good. Memory isn't the only option: in two companies, raters who kept structured diaries recalled more and gave less inflated ratings (DeNisi and Peters).

If you're an employee. You can't fix the room, but you can make it easier to see you. Log results, not tasks: what changed, by how much, and who relied on it. Show the log to your manager every quarter. Write down the quiet work too, because nobody credits what they can't see. This isn't about getting louder. It's about getting legible.

So ask it plainly: are we rewarding the people who create the most value, or the people whose value is easiest for us to see?