In both cases in the article it's almost certainly down to "stop word" removal. The titles contain "the" and "and". These kind of words get filtered out becau

In both cases in the article it's almost certainly down to "stop word" removal. The titles contain "the" and "and". These kind of words get filtered out because they make indexing incredibly difficult, ie. you search for "the" and get every single book containing "the" in the title (and sometimes description, depending on how deep the search is). So they exclude extremely common words to get eg. "catch kill". Then they have to rank the results for "catch kill" by some other heuristic - like popularity - and you get a list of books that people weren't looking for. Not to mention the first two books in the list were actually exact matches, so maybe they were more popular on the site. Big search engines do try to correct for this, but it's not an easy problem. Usually results in some hard-coded exceptions for names like "The Who", but it's not trivially scalable. In the case of your google maps results, I imagine it'll be something like (popularity * distance), which isn't necessarily the best heuristic. The problem is that you can't please everyone with heuristics. Give power users the option to adjust the search to distance-first, popularity-first or any other of a dozen different combos and you end up with a confusing mess of a UI. Simplify it and give them your nicely hand-crafted best guess and there will always be someone complaining...