Initially, things look disjointed. Ars notes that "a game with an 80 on Metacritic ... is just as likely to bring in about 56,000 estimated Steam sales (at the 25th percentile) as it is to earn 375,000 estimated Steam sales (at the 75th percentile)." However, looking at the median sales and aggregating Metacritic scores into 10-point ranges, a more identifiable trend emerges. According to Ars' estimates, a game with a median Metacritic score of 90 or higher will sell 50 times the amount of copies that a game with a Metacritic of 30 or lower sells.
Some caveats: Ars Technica's data represents estimates based on publicly-available information, and Metacritic itself weighs reviews from different sites differently. One man's trash may be another man's treasure, but Metacritic may be inclined to agree with one side more than the other. Also, the data analyzed doesn't take into account discount events such as the Steam sales - which we all know are the bank account's natural enemy - so there are additional factors that could be at play here. Lastly, Ars notes that individual games may buck the trend: Orion: Dino Horde has a Metacritic rating of 36, yet an estimated 314,000 Steam sales, while NBA 2K13, despite its rating of 90 on Metacritic, has only sold an estimated 50,000 copies on Steam.
Nonetheless, there appears to be some sort of correlation between higher Metacritic ratings and higher sales on Steam. Causation, though? Maybe not so much. And while this information might seem like common sense, now there are numbers to back up your intuitions.