Chad Perrin: SOB

13 June 2007

The software market for lemons, and how open source software fits.

Filed under: Geek,Liberty — apotheon @ 12:17

There’s a common meme floating around the so-called “blogosphere” that relates economist George Akerlof’s Nobel-winning “market for lemons” ideas to the software industry:

  • David N. Welton‘s Web Hosting – A Market For Lemons makes the case for the lemon market phenomenon applying to the web services hosting market. It hearkens back to 2005, but it has resurfaced in references from other sources because of the current “market for lemons” meme propagation.
  • Bruce Shneier‘s A Security Market for Lemons relates the lemon market phenomenon to the often abysmal quality of computer security products. Shneier may be largely to blame for the current popularity of the meme.
  • Reg Braithwaite‘s The Not So Big Software Design, where he tackles both the software engineer employment market and the cutom software development market. If you’ve ever wondered why so many huge corporations use cruddy customized ERP “solutions” that literally cost millions of dollars, this might give you a hint. Of more immediate concern to most of us, this also pretty well illustrates how it might be that we ended up with employers hiring all the wrong people and refusing to pay the right people what they’re worth without just calling hiring managers idiots — though he mostly lays that subject at the feet of “steveblgh” at reddit and focuses on the custom software subject.
  • Richard Tibbetts ties some of this together quite well in The Lemons Meme in Software, and brings up resale value — which actually further devalues custom ERP packages and the like.

The short version of the theory of lemon markets goes something like this:

  1. Customers can’t tell just by looking at a car whether it’s any good.
  2. Sellers know more about the quality of what they’re selling than the customers.
  3. Sellers with crap products are likely to lie about the quality of the products when they can get away with it.
  4. Customers tend to become wary of high-priced products, because if the car turns out to be a lemon they’ve lost more money than with a lower-priced product.
  5. Nobody pays more for a car than within spitting distance of the average quality of car in that product’s class.
  6. Higher quality cars that aren’t worth selling at the depreciated “average quality” price are thus priced out of the market.
  7. Without those higher quality cars dragging up the curve, the average quality of cars on the market drops — as does the average price.
  8. Wash. Rinse. Repeat.

Thanks to trends in the proprietary software industry over the last twenty years or so, the tendency is for the customer to have less and less information about the software he or she considers buying. Software of all stripes is affected by the lemon market phenomenon, but unlike in other (only slightly) healthier lemon markets, there’s a lower bound on price because of the effective commercial monopoly on low-end general purpose computing software. This means that the lemon market phenomenon can cause quality to drop precipitously while ensuring that prices do not follow suit.

The lemon market phenomenon, in this case, essentially drags customers out of the high-end software range and into the mediocre-end software range, because they’re unwilling to pay the price for high-end software but unwilling to drop below the mediocre quality of the software they find themselves using as a result. Things are changing, however.

Open source software development, because of its apparently unique relationship to an information market, throws a spanner in the works of the lemon market. Think about it: the center of gravity for the software industry is information management. There’s a reason they call this stuff “Information Technology”, after all. Software itself is just information — but in attempts to shore up the defenses of positions of market dominance, the proprietary software vendors have embarked on a crusade against the sharing of information related to their own software offerings. They wish to maintain asymmetrical information access so that they are in a position of power with regard to the customers. Open source software, by contrast, predicates the success of its business model (for lack of a better term) on the free, unfettered sharing of information.

Because the lemon market phenomenon is entirely dependent upon asymmetry, the rise of open source software in the public consciousness is beginning to shake up the software industry. The lemon market effect that keeps MS Windows customers (I mean businesses as much as, if not more than, end users) from going back to Unix and mainframe systems is being undermined by the simple facts that:

  1. No lemon market can survive roughly symmetrical information access.
  2. Open source software, such as free unices, is available at a price that Microsoft just can’t match under its current business model.

I’ve talked before this about the fact that the combination of the “wild west” aspects of the Internet and the freedoms available through open source software leads to a purer free market economy in practice than just about anything else we’ve seen. The monkeywrench in the lemon market phenomenon that has ruled much of the software industry for so long is a significant part of the reason for that.

Computer security, web hosting, and custom software development are really just special cases of the overall lemon market tendencies of a closed source software industry predicated upon notions like security through obscurity, astroturfing, and intellectual protectionism. Principles of security through visibility, social networking, and a protected public domain serve as foils for the lemon market effect in the software industry, however. We’re on the right track. Just don’t let us get derailed.

6 Comments

  1. the cure for any lemon situation is better information. we simply need better ways to gather and share it.

    this is a project i’ve been putting a lot of thought into.

    Comment by sosiouxme — 13 June 2007 @ 08:46

  2. Now you’ve got me curious. What kind of project have you got in mind?

    Comment by apotheon — 13 June 2007 @ 09:45

  3. I think this meme totally misunderstands and misapplies the economic theory of a “lemon market”.

    Lemon markets aren’t just markets where the seller knows a bit more than the buyer: that’s true of practically every consumer good.

    A used car market MAY be a lemon market, because the buyer has no way of knowing which individual car is a lemon or not.

    A new car market is not a lemon market, even though the seller may well know more about the cars. The buyer here can still get plenty of information about which models are better than others: from friends, from consumer reports, even from the marketing.

    Even more with software products: every copy of, say “Symantec AntiVirus Corporate Edition” is EXACTLY THE SAME as every other. These means by reading consumer reports and getting advice, it’s possible to get pretty good information on the quality of it.

    Furthermore, a lemon market depends on sellers with a high-quality item withdrawing it from the market. If you have a reliable used car, you might sell it or give it away to friends and family who know it’s true value.

    I haven’t seen any explanation of how and why software manufacturers supposedly withdraw their high-value products from the market. How are they selling their product? Software is infinitely replicable: why can’t they just sell extra copies on the market too?

    The people using this meme don’t seem to understand the basic concept at all. They seem to be skim-reading Wikipedia, thinking “oh, there’s some kind of economic thingy where everything in a market is crap” and then deciding this must apply to some kind of business they’re not happy with.

    Comment by R Mutt — 14 June 2007 @ 09:24

  4. Lemon markets aren’t just markets where the seller knows a bit more than the buyer: that’s true of practically every consumer good.

    Of course they aren’t. They’re markets where the seller knows a lot more than the buyer, and as a result of a number of consequences of that fact there’s increasing pressure on the market to provide more lemons and fewer good products. I don’t think anyone’s saying that any time a seller knows more than a buyer you automatically get a lemon market.

    Even more with software products: every copy of, say “Symantec AntiVirus Corporate Edition” is EXACTLY THE SAME as every other. These means by reading consumer reports and getting advice, it’s possible to get pretty good information on the quality of it.

    Yes and no. The mechanisms for the market pressure to provide lower-quality products are different than for a used car lemon market, but they do still exist. The problem is that end users can’t tell the difference between differing levels of quality with software much of the time — even when those end users are presumably knowledgeable tech journalists and IT professionals. Since these end users of every stripe make up the body of reviewers, reputation doesn’t count for much when it comes to quality a lot of the time: people think they know what quality is because they just don’t know any better, so they tell other people what they think is high quality software, and as a result a bunch of crap software with good marketing (like Symantec security software) gets all the business.

    I haven’t seen any explanation of how and why software manufacturers supposedly withdraw their high-value products from the market. How are they selling their product? Software is infinitely replicable: why can’t they just sell extra copies on the market too?

    Since the market is largely dominated by huge corporations, the high quality software that is getting priced out of the market is the stuff produced by small companies, and the high quality software the larger corporations aren’t producing because it’s not cost effective when they can make more money spending less on quality development.

    Sure, it’s possible to figure out what software is good and what software is bad, but you have to think outside the standard software industry box to do that. Good luck.

    The people using this meme don’t seem to understand the basic concept at all. They seem to be skim-reading Wikipedia, thinking “oh, there’s some kind of economic thingy where everything in a market is crap” and then deciding this must apply to some kind of business they’re not happy with.

    While that may be true of some, it’s certainly not true of all. Reg Braithwaite and Bruce Schneier, at least, seem to know what they’re talking about. I think part of the problem is that they (and I, for that matter) have tended to assume more ability for readers to figure out some of the translations of concepts between a used car market and a software market, and how similar economic effects can be had with a somewhat different business model, than should really be assumed.

    Comment by apotheon — 14 June 2007 @ 10:55

  5. What kind of project have you got in mind?

    i can’t tell you; it’s going to make me rich! heheh. well, ok.

    people love to share their experiences with products, whether lemons or lemonade. and whenever someone is considering a product, they frequently ask around for recommendations. there’s an incredible amount of valuable information (though of dubious reliability) in this plain gossip. just no one’s found a way to capture it, or for the most part even considered the possibility.

    but increasingly this gossip is already online. it’s already being written down and emailed/IMed/posted. so i’ve been thinking about ways to channel questions and opinions, turn them into quantified data, and organize that in a way that’s universally useable. there are all these disparate sites now that collect opinions for a narrow subject field, with narrow criteria. i want a site where you can search for or request opinions you trust on everything, with any criteria you can think of. it won’t do much to detect individual lemon cars, but it should help with anything that can be experienced in common.

    that’s it in a nutshell. i need to work on my elevator pitch. and my delusions of grandeur i suppose…

    Comment by sosiouxme — 14 June 2007 @ 10:15

  6. Excellent! That sounds like it could be fun — and, of course, challenging — to get running.

    Comment by apotheon — 14 June 2007 @ 11:29

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

All original content Copyright Chad Perrin: Distributed under the terms of the Open Works License