The Short List Problem

Ask anybody for a recommendation and they will offer several, because offering several feels more helpful and is enormously less useful. The list of ten is a way of not answering. It looks like generosity and functions as a refusal to choose, and the person who asked walks away with more work than they arrived with.

There is a reason the instinct is so strong. Choosing one is a bet. The list of ten spreads the bet across ten options, so no single failure attaches to the recommender. If the listener picks the fourth item and hates it, well, they picked it. The list is a way of enjoying the credit of good taste while transferring the risk to whoever asked.

The mathematics of the situation are not kind either. Every additional option lowers the chance that any of them are acted upon, because the cost of deciding rises faster than the value of having more choices. Beyond about three items, the list is functionally empty. It is a display of range, and range was never what anybody wanted.

The counter-argument is that people are different and one recommendation cannot fit everyone. True, and irrelevant to the case where a specific person has asked a specific question. What they are asking for is not coverage. They are asking to borrow somebody else's judgement for a moment so they can stop exercising their own, which is exhausting and which they have been doing all day.

The one-item answer also does something the list cannot: it makes the recommender accountable. If the single thing lands badly, that is visible, and it should be. That visibility is what makes the next single thing worth listening to. A track record only accumulates if the individual bets are legible, and the list of ten is specifically designed to keep them illegible.

In practice the compromise that works is one item with one alternative, and the alternative stated as a fallback rather than a peer. Start here. If the first fifteen minutes do nothing, this other one instead. That is still a decision, still accountable, and it acknowledges the possibility of a miss without diluting the primary bet into a menu.

There is also the matter of what a list does to the things on it. Placed in a row of ten, every item loses its particular reason for existing and becomes an interchangeable unit of content. The reason is the thing being recommended, more than the object. Ten reasons in a row cancel each other out; nobody retains the fourth one.

So the format holds at one. One thing, one reason, one week, said plainly. Everything else that was good this week can wait, and most of it will turn out not to have been good enough to wait for, which is itself the useful information.