I went back to the actual Vickers paper rather than relying on the abstract, because my previous wording was a bit too strong.
The 2010 AES paper is Earl Vickers, The Loudness War: Background, Speculation, and Recommendations. The 2011 JAES article Do Louder, Hypercompressed Recordings Sell Better? is a shorter, slightly updated version.
The conclusion is not simply “loudness doesn’t correlate with sales.” It’s more nuanced—and actually more interesting.
What he found
Vickers looked at commercially successful recordings and compared loudness/dynamic-range measures with their sales/chart performance. His conclusion was that among albums that reached the Billboard Top 200, he found no evidence that greater hypercompression correlated with better sales performance.
Vickers himself later clarified this in the AES discussion:
“I don’t see evidence of a correlation between hypercompression and better sales among the albums that reached the Billboard top 200”
But he immediately adds an important qualification:
“the albums that charted did indeed tend to be more compressed than those that did not.”
That’s important. His data therefore cannot tell us that compression had no commercial effect. There is a selection problem: perhaps compression helped records get noticed/chart in the first place, while further compression didn’t determine how well they subsequently performed. Or perhaps successful commercial genres simply happened to be mastered louder. Correlation can’t distinguish those explanations.
Listener preference
He also examines the assumption that people actually prefer hypercompressed recordings. The evidence he reviews doesn’t support the simplistic idea that more compression → greater preference.
This is subtly different from the familiar loudness effect. If you present essentially the same material at different playback levels without matching them, people can favor the louder presentation. But that does not demonstrate that listeners prefer the distortion/reduced dynamics necessary to make a recording intrinsically louder.
That’s one of the conceptual mistakes Vickers is attacking.
And his game-theory argument is important
His explanation of how the loudness war could persist without evidence that it increases sales is essentially a competitive arms race.
Imagine every label thinks:
If everyone else’s record is loud and mine isn’t, mine might sound weak by comparison.
Then making your record loud becomes individually rational even if nobody has demonstrated that the collective escalation benefits anyone. Once competitors do it, the perceived cost of not participating can be sufficient to keep the process going.
That’s why his 2010 abstract carefully says:
“Evidence is presented to question the assumption that loudness is significantly correlated to listener preference and sales rankings.”
It doesn’t say “we demonstrate that loudness has no effect on sales.”
And Vickers himself later said in response to discussion of an aggressively compressed but highly rated radio station:
“It’s hard to know whether overcompression has any relationship to that station’s popularity.”
He suspected the phenomenon was driven substantially by assumptions and what Bob Katz called “loudness envy.”
So, correcting what I said earlier
My earlier:
“research … found no significant relationship between loudness and sales ranking”
made the evidence sound stronger than it is.
A more accurate statement is:
Vickers found no evidence that greater hypercompression predicted better sales ranking among albums that already reached the Billboard Top 200. Interestingly, charting albums were more compressed than non-charting albums. His study therefore neither proves that compression helped sales nor that it didn’t.
And that means the answer to your earlier incredulous question really is remarkably close to yes:
After decades of loudness war, we apparently still don’t have good causal evidence showing that hypercompressing a recording makes people buy more of it.
We know the industry believed it could provide a competitive advantage. We know louder presentation can create immediate perceptual impact. We know charting records tended to be more compressed. But the causal chain
more compression → greater consumer preference → greater sales
was apparently never established.
Which makes Californication a particularly bad example for anyone trying to infer consumer preference from sales: its enormous sales cannot tell us whether the mastering helped, hurt, or made essentially no difference to those sales.
