Pattern matching at it's most basic is about finding things that seem to fit together.
In venture capital that often means looking at startups and interpreting what works well, or what makes a good entrepreneur. There are plenty of problems with the method - Chris Dixon wrote about a few of them in 2012, and there is significant research that suggests that pattern matching through things like gender and race may unconsciously lead to much narrower ideas of what works, and poorer outcomes for picking winners.
But one of the things that's not dug out as often is that looking for patterns that don't match is a valuable skill. Another way of thinking about it is: everyone wants to know what makes a winner, but not enough people want to know why something doesn't work, why a pattern you expected to fit, simply doesn't. As it turns out, those tend to be the underpinning to finding patterns that create growth in a company (and probably as an individual, too).
For example, one of the most complex questions in an early stage business is: we've found something that works, but how do we know it will last? You can pattern match your way to more users or customers ("hey this works let's do more of it"), but unless you're paying attention to which patterns don't match, you'll likely get hit hard when a tactic doesn't work. There's a very small peek at that struggle in a recent update from Mike Wilner at Compass, which matches designers/developers to small business owners & entrepreneurs who need a smart, basic website.
Both sides of the coin are important. You have to be willing to seek things that don't make sense, because they can help you hedge against your own unconscious bias. And it increases your ability to see things at a much larger scale, instead of simply chasing things that fit what you've already run into.
