I keep running into the same addlepated thought, here and in other forums. It goes something like this:
Even here at ASR, I've seen a few (just a few) posters go down the tubes trying to defend the idea. On Audiogon it's practically gospel from some of the most prolific posters. So let me just lay out here why it is nonsense, and I can link to it if it comes up again.
Let me state my claim explicitly (in after-posting edit):
Market success proves neither a)better fidelity/accuracy, b) better sound quality, nor c) better engineering quality. It proves only that a product met the revealed preferences of some buyers, which may include, but are not limited to those three criteria. In real life, there are many criteria that complement, confuse, or even obscure these functional sound criteria. It is fine to use inaudible criteria, but sloppily confusing them with sound quality, or, worse, claiming market success *must correspond* to audible quality, only muddies the waters for those of us who prioritize fidelity and sound quality.
I am not anti-market, in fact I've been called a "market fundamentalist" by some. I make my living in financial markets. Markets allocate capital to societal preferences and capacity far better than top-down planning can. But look at the way I stated it: "preferences", "capacity", "better than top-down planning". Markets are not all-knowing, they are not moral judges, and they reflect preferences that can be self-destructive. Consider, for instance, lottery tickets, which face staggering demand for objectively minuscule expected value. There are a few bona fide instances of market failure, and even more market outcomes that are not, for example, healthy. If you want to be successful in financial markets, you need to understand cognitive bias, which we will discuss here momentarily. Actually, if these biases/behavioral irrationalities didn't exist, there would be no arbitrage profits in markets.
What do we mean by a market success? It means that people made enough purchases for the “successful” seller to survive. In audio we often don't know if they are making a decent profit, but let's assume,
arguendo, that they are. So we say someone has achieved market success if they are making money selling a type of audio gear. That means that customers are, in a sense, voting for that gear,
on a dollar-profit-weighted basis. Success can come from making an insanely profitable sale to one or two people, or it might come from a less profitable sale to a lot of people. Margin or volume, or both.
So there's the first flaw in the theory - a small number of high margin buyers can keep a company afloat. At the very high-priced end, If two dribbling morons give a vendor a 10,000% profit margin on five figure
liquid metal cable and a custom burn-in device, the vendor may make a profit and declare himself a success. At the lower-cost end, where many tweaks reside, the lower the cost of the component or tweak, the less dollar volume it takes to “succeed”. Like
these, for instance. There's hardly any fixed overhead or much marginal product cost associated with some tweaks, and the vendor can survive indefinitely with very few customers. One of the tweaks on that list involves only a phone call, thus it has almost zero marginal cost. These kinds of low-volume, low cost transactions are insufficient to generalize about customer preference and utility.
The next thing we have to discuss is 'value', or perhaps
utility (in the economics sense):
Notice that utility in the examples comes from not just how strong the deadbolt is, or the (sound?) quality of the movie in some dimension, but from any aspect of customer satisfaction (of want). Preferences, or each buyer's definition of "value', reflect utility
outside of strictly functional qualities. For the purposes of this discussion, we can use "value" and "utility" interchangeably. Proponents of the argument stated at the outset tend to use "value". The point here isn't the term, but the fact that value comprises multiple criteria (conscious and unconscious) that are not strictly
audible sonic quality. They may also include not just
engineering (non-audible) quality, at least the kind that we can measure.
I submit that high end buyers have all kinds of inaudible and non-engineering wants to satisfy. Even more important, we are only barely aware of some of them when we spend our money. Here are a few:
Signaling
Signaling is simply sending information without stating or explaining it. Signaling is ubiquitous in human interaction - we signal status, we signal our knowledge, we signal our job fitness or academic achievements and we signal our wealth.
Purchasing luxury, or very expensive, goods is a way to signal wealth and status. People do it with cars, watches, clothes, houses, art, and even philanthropy. That big wristwatch says you are wealthy enough to afford it. It certainly doesn't tell time better, nor is it necessarily more visible to your presbyopic eyes.
Do you think your college education -strictly what you *learned* in college - was worth the tuition? Would it be worth GWU's tuition of $57,500? Is GWU? One of the most fascinating
books I've read recently suggests that college isn’t about learning at all. I went to Yale, but I don't think the *education* there is any better than a number of much cheaper colleges. I haven't found too many people who would argue that. So what are we paying for? Networking, signaling to employers, status, fancy facilities, access to recruiting, sports opporunities...the list goes on.
Robin Hanson, of GMU, writes a lot about signaling. One of the most pernicious signaling effects in this day and age is "
inward apology". This is when you stick your neck out to believe something ridiculous in order to signal your loyalty to a movement or group:
I bet any of you can think of an example. Conspiracy theories are a big one. Think about one aspect of this: Inward apologetics must be
costly to be effective signaling. Like blowing a lot of money on cable risers, perhaps? Or doubling down on internet forums about quantum stabilizers? If you were a dealer, might you engage in this behavior? If you had spent years building an objectively inferior system and becoming a big shot on Audiogon, would you do this? "
Out of the crooked timber of humanity, no straight thing was ever made "
Revealed Preference
We can't discuss signaling without revealed preference. Know of anyone who flies a private jet but can't stop preaching about global warming? Do you think people virtue signal, while being hypocrites in their personal life? This is revealed preference - when your actions betray a different set of priorities than your words. Audiophiles try to conceal their revealed preference by inventing nonsense or non-descriptive sonic characteristics like Pace and Rhythm and Timing, as distinct from measurable rise times and frequency response. They say they care about "resolution" but buy an objectively concealing low S/N DAC.
Looks
Here the intersection with wristwatches and cars (also predominantly male hobbies) is strong. Look hard at the Primaluna amplifier below. The measurements are bad, and suggest unpredictable performance, particularly with variable loads. But that Italian design, oh my! Can you tell me that they, or any tube amp manufacturer, aren't selling the visual appeal of glowing tubes? A little tech-meets-modern, or steampunk.
I just asked my wife to describe this style, and the first words she said were "phallic brutalism". Well, it is a male-dominated hobby.
View attachment 36558
It isn't just that buyers want something that looks good, they want something that loudly signals their hobby and the status in the hobby they believe they have achieved. They want their equipment to scream "I have esoteric but refined tastes". Certainly there's also lots of evidence that staring at an impressive object like this will "
prime" the viewer into appreciating the sound, the other parts of the experience.
Having mentioned "priming", we should consider some other
behavioral finance effects. <--that link will take you to a nice summary with a top-ten list of known behavioral irrationalities. You can hover over each item on the list to see a short description. Many of these effects were identified in experiments with large numbers of market/competitive transactions, and have also been observed in real world transactions. Daniel Kahneman and Amos Tversky's nobel prize is based on
Prospect Theory, which is a sort of synthesis of these effects. They are real and pernicious, and awareness does not necessarily make them go away.
The most relevant of these for High End Audio, I think, are as follows:
Narrative fallacy (I realize I have this one - well, I have all of them to some extent, but I think this one is strong with me) - we want to believe a good story. That lonely guy in his garage fiddling with MOSFETs has invented something that completely blows the big brands out of the water. Forget all that expensive Harmon research. Those wacky rebels at PS audio fiddled around with FPGAs and just totally solved all the alleged audible problems with digital audio, despite decades of well-funded research by specialists. And they keep doing it every six months - "night and day". We want to believe these kinds of outsider revolutions a) happen more often than they do and b) we are just the highly discriminating people to have discovered them. We are back to signaling, then. The narrative says something good about us.
Confirmation Bias. This is a bit like priming. After you read all those great subjective reviews, and especially if you liked the looks or the story, you are ready to hear the equipment. SURPRISE, it sounds awesome. Well....you were primed.
Loss Aversion and decision regret. I think this is more subtle, and I'm using the terms in a slightly different way from the literature. Even if you realize you made a mistake buying snake oil in the past, and its been proved to you via measurements and possibly a blind test, you bargain with those empirical results so your ego doesn't take/admit a loss. "Blind tests are bogus", you say "I can definitely hear a difference. "Not everything audible can be measured". Perhaps you let the vendor suck you into his quantum theory if you don't already know enough physics to know you are being conned/soothed. You keep coming back to the trough - to do otherwise would hurt. You might also call this the
sunk cost fallacy.
Overconfidence, self-attribution bias, and herding are also clearly problems. It sounded great to you because of your skills in finding the right equipment synergies! You have golden ears, and more refined tastes than the masses! Michael Fremer raved about it in Stereophile and a bunch of people on Audioholics say it's great, so it must be. If you find yourself thinking these things (to paraphrase Jeff Buckley)...
you may be irrational.
Two last things about market transactions: First,
Asymmetrical Information. All markets involve information asymmetries, and when they are severe (such as medicine), market outcomes are heavily distorted. The fact is that the average equipment manufacturer knows way more about her equipment than you do. They have measurements they withheld. Only they know the actual results of their own blind testing, if they did any. They listen to equipment all day. They aren't telling you the pressure they are under to differentiate their product. It is extremely time-consuming to learn what you need to learn to make a perfectly rational purchase. In fact it is nearly impossible. You can't listen to everything. So you rely on heuristics, which usually encode some kind of bias. Information on audio is increasingly available and transparent, but there are still significant asymmetries, especially for us non-engineers. I just want good sound, I don't want to study acoustics (well, maybe a little).
Market Fragmentation increases the information asymmetry problem. We can only listen to a small sample of what's available. There are fewer and fewer audio stores, and the dealer can
frame/anchor your decision within his slim coverage. You can order stuff on the internet and return it, but if you order some JBL Everests, you might have a little trouble returning them. High end speakers typically weigh 100 pounds or more. And the industry has bamboozled you with talk of "burn-in" to get you past the return period.
Finally, it is clear that audio buyers often value build quality and engineering, even when the effect on audibility is negligible or zero. A quick run through any audio mag will tell you we love quality knobs, interconnects and panels (back to 'phallic brutalism'). Audio magazines practically have centerfolds. It seems we love the appearance so much we desperately want those things to mean they
sound better. We all know, in our mind of minds, that it usually ain't so.
So put this all together. Market equilibrium is where the downward-sloping demand (price=X, quantity=Y) curve meets the upward sloping supply curve for some bundle of transactional attributes (in reality this is always dynamically adjusting). We've shown that this bundle of attributes includes both real qualities and irrational preferences that are well beyond audible sonic quality. We've also shown that occasionally this equilibrium reflects only a very small subset of individuals, choice is artificially restricted, or information asymmetry distorts the equilibrium. In these cases, the result can't validly be generalized to a wider group. In fact, some might even argue that high end audio is a
Giffen Good, an extreme case of signaling that defies equilibrium altogether. For these reasons, the market success, at equilibrium, of a given product cannot be used as an argument for, let alone proof of, audible sonic quality, or engineering quality.
By the way, knowledge of your biases can only take you so far. No discussion of this topic is complete without
Sarah Lichtenstein's "Money Pump", in which she explains to people *how* she is ripping them off using another well-known bias, and they...just keep going for it. There's a delightful audio tape of her doing this, which I'll have to re-find, it's buried in the interwebs. (transcript
here). Other examples of conscious irrationality or conscious cognitive/moral dissonance include the
Stanford/Zimbardo Prison Experiment, or the Y
ale/Milgram torture experiments. Humans!
If you haven't already realized it, this post also constitutes a thorough argument for
blind testing, as the only way to strip the sonic experience of its visual, status, confirmation bias, and other baggage. You will, however, be able to signal your objectivism if the cheap dongle happens to win. There's probably a bit of that in objective circles.
)
