Skip to main content

In The Streaming World, How Much Do Hits Matter?

A lot. A tiny percentage of songs now accounts for a large percentage of total plays.

For years, media and technology executives have been saying the age of the blockbuster was over. Now that online consumers can choose from among the tens of millions of tracks available online, hits would inevitably become less important relative to the sheer amount of music in what author Chris Anderson termed “the long tail” in 2004. Hits, wrote Anderson, would no longer be “quite the economic force they once were.”

The reality of the music streaming market is very different. Last year, Drake scored 6.16 billion on-demand audio streams in the United States, according to MRC Data — more than any other artist, and 0.7% of the 877.2 billion total. That exceeds the 4.74 billion streams generated by the 53.69 million tracks that were streamed fewer than 1,000 times each.


For all the changes in the music business — now dominated by digital over physical and streaming rather than sales — hits remain as important as ever. In 2020, nearly half of the 877.2 billion on-demand audio streams in the United States came from just 13,521 songs that were streamed over 10 million times — or 0.022% of the titles that MRC Data tracked.

To get a sense of whether, or how much, hits dominate the business, Billboard used an MRC Data analysis of on-demand streaming to separate the 2020 market into five “buckets”: songs streamed over 10 million times; between 1 million and 10 million times; between 50,000 and 1 million times; between 100 and 50,000 times; and fewer than 100 times. To compare the industry today with that of 25 years ago, Billboard then examined a 1995 report from MRC Data’s predecessor company name, Nielsen SoundScan, which separated that year’s sales into albums that sold over 250,000 copies; between 25,000 and 250,000 copies; between 5,000 and 25,000 copies; and fewer than 5,000 copies.

Courtesy Photo

Hits remain important by any measure. Combined, the first two buckets of songs streamed over 1 million times — which includes 96,779 tracks, or just 0.16% of the 61.2 million available — accounted for slightly more than three-quarters of total on-demand streaming. On the other end of the spectrum, even when discounting the 68.72% of available tracks that were streamed fewer than 100 times, many of which were presumably uploaded by hobbyists, the 18.26 million songs streamed between 100 and 50,000 times — the much-vaunted long tail that accounts for 29.8% of tracks — only account for 6.25% of total listening. Culturally, the easy accessibility of digital distribution represents a revolution. In music, though, it’s just not much of a business.

Market concentration was similar in 2018 and 2019 for the various stream buckets, with a variable swing of no more than .5%. Likewise, the number of titles in each bucket yielded similar percentages in 2018 and 2019 with a spread differential of no more than 20 basis points, or .2%. However, the bottom bucket that had streams between 1 and 100 plays — the bucket containing the long tail with the most titles — as a percentage of overall titles, grew by almost two percentage points between 2018 and 2020, starting out at 66.9% of all titles in 2019; growing slightly to 67.1% in 2019 and then to  68.7% of 2020’s total tiles.

Back in 1995, the 336 albums that sold over 250,000 copies — 0.2% of the nearly 147,000 releases available at the time — generated almost 40% of sales. Combined with the second bucket of albums that sold between 25,000 and 250,000 copies, that meant 3,328 albums, or 2.2% of releases, accounted for 72.5% of total sales.

At least part of the reason such a small percentage of songs now accounts for such a large percentage of total streaming is that there’s so much more music available than there was then. While consumers in 1995 could choose among 146,693 albums — which, combined, probably contained somewhere between 1.5 million and 1.75 million songs — in 2020, MRC Data tracked 61,189,195 songs, the equivalent of about 5 million to 6 million albums. But the business of unpopular music isn’t exactly exploding. In 1995, the 91.3% of releases that sold fewer than 5,000 copies accounted for 11% of sales, while in 2020, the 98.56% of releases that were streamed fewer than 50,000 times made up 6.34% of streaming consumption.

The idea that the long tail represents a significant business opportunity doesn’t hold up, either — at least when it comes to music streaming. In 1995, the bottom bucket of 134,000 albums that sold fewer than 5,000 copies each would have accounted for 67 million in total sales and brought labels about $536 million in revenue, assuming an average wholesale price of $8 per album. Last year, however, the 60.3 million songs in the bottom two buckets — which were streamed fewer than 50,000 times each and accounted for 98.6% of available tracks — accounted for a combined total of 55.64 billion streams, or 6.43% of the total, and only generated $295 million in revenue, assuming a blended per-stream rate of $0.0053. In other words, even before factoring in inflation, the business of unpopular music is just under half the size it was a quarter of a century ago. Turns out hits still have some life in them after all.

A version of this story will appear in the Dec. 18, 2021 issue of Billboard.