In a study published in Administrative Science Quarterly, Berg tested his hypothesis by looking at a data set of more than 3 million songs by almost 60,000 artists from 1959 to 2010. Of those artists, 4,857 had one or more hit songs.
To determine the creativity of songs, Berg used an algorithm to measure features such as key, tempo, time signature and danceability. He also calculated how varied an artist’s portfolio was each year they released one or more songs. What Berg ultimately found was that artists who had a novel or varied portfolio before an initial hit were more likely to continue creating hits, but a novel portfolio was less likely to yield the initial hit.
“Novelty is a double-edged sword,” Berg tells the Atlantic’s Derek Thompson. “Being very different from the mainstream is really, really bad for your likelihood of initially making a hit when you’re not well known. But once you have a hit, novelty suddenly becomes a huge asset that is likely to sustain your success.”
Artists are also more likely to succeed if they create new songs that are related to their existing work, Berg explains in a statement, but they also have to balance keeping up with current trends.