Creutz notes that scores can be "somewhat predictive" of a game's success, but are "unlikely in and of themselves to drive or undermine the success of a game." He then mentions the industry scuttlebutt that many publishers work very hard to jockey scores, saying that publishers' time would be better spent on development instead of "grade-grubbing."
The major factors in purchasing decisions, according to Creutz, are genre, whether a player liked the previous game in a series and, unsurprisingly, price. So there you have it publishers, to be successful, all you have to do is make a sequel to a popular game in a genre that a lot of people like. And make it cheap. If that's not possible, two out of three ain't bad.