The Pareto principle came from the observation it's common to find distributions where 80% of the effects come from 20% of the causes – 80% of land in Italy is owned by 20% of people, 80% of book sales are from 20% of authors, and so on.
By now the principle is well known, and we're often advised to find which 20% we should be focusing on that would give us 80% of our results – which 20% of features would fit the needs to 80% of our users, which 20% of our projects would have 80% of the impact, which 20% of investments would product 80% of the profits, and so on.
But we need to have a more nuanced view. Identifying the right 20% beforehand can be difficult. After-the-fact, sure, you know which 20% of your investments produced 80% of your profits. But beforehand? It's not always easy.
In the startup world, investors know a small number of their investments will drive most of their returns, but they can't predict which ones. So they diversify their investments across a large number of startups, hoping a few will be incredibly successful and pay for the majority that won't.