And here's an interesting perspective on the value of search that isn't too relevant. Michael Shrage of MIT argues that when we're doing the kinds of learning and discovery searches for research that we often do, our notion of what's relevant in search results will expand and change as we go through each iteration of a search. We may go into a search with an idea of the kinds of results we want to get, but that idea will shift with our first results, which can tell us that maybe we want to go off on a tangent we hadn't considered.
Of course, if we're looking for something like a particular software review or the location of a store, we want highly relevant results. But many of our searches are research-based and greater creativity and innovation will be sparked by having search results that may initially seem less relevant. In other words, we need to leave some room for "irrelevancy" in the search process because we don't always know going into a search what will really be relevant to our initial research question.