In my first post, I said that I probably wouldn’t be posting
any treatises on Economics. Well, I’m sorry to say, I’m posting one now. Why?
Because yesterday the O'Reilly Network
Developer News, pointed to an article
written by David Stutz. For those who don’t recall, David Stutz headed up the
Rotor project before leaving Microsoft, and was a vocal advocate of open source
both before joining Microsoft and when he was employed at Microsoft. His parting open letter
to Microsoft, written a little more than a year ago, found some cachet within
open source circles.
My only personal contact with David occurred just before he
left Microsoft. The lead program manager for Mac Word, Jason Webber, was
leaving Mac BU to join the Visual Studio.NET team. One of us had the brilliant
idea of sending Jason off with the, then yet to be released, version of Rotor
that ran under Mac OS X, so I fired off a piece of e-mail to David asking where
we might find the bits. David was more than accommodating, and very easygoing.
I’m inclined to think that David’s really a very nice guy,
which is why I’m so very disappointed with his whole concept of software “commodification”.
Besides the fact that “commodification” is the wrong word (the correct word is
“commoditization”), he really fails to grasp the essential details of the
concept. As a result, his conclusions are misguided at best and procrustean at
worst. Hence, this post.
If you’ve slogged your way through David’s essay and found
yourself saying, “Huh?” don’t feel bad. He’s talking about the process by
which certain goods and/or services become “commodities”. As David correctly
points out, commodities are fungible regardless of the producer. You can
substitute one producer’s output for another producer’s output without noticing
the difference. As such, commodity producers tend to compete solely on the
basis of price, and relative profit margins are determined by each individual
producer’s costs of production (what accountants call “cost of goods sold” or
COGS). Commodity markets tend to most closely resemble what Economists call “perfect
For example, milk is a commodity. Given a taste test
between Lucerne 2% milk and Wilcox 2% milk, only a very small percentage of the
population would be able to tell the difference. Between various stores, you
might find as much as a 50% difference in prices for a gallon of 2% milk, but
that’s mostly because some retail sellers use milk as a loss-leader to bring
people into their stores.
Contrast this with, say, DVD players. If you go to PC World’s
grabber, you’ll find the most expensive player priced at just under
$4100.00. The least expensive DVD player is $49.00. With a price differential
of nearly two orders of magnitude, clearly the market for DVD players is not a
commodity market. If we’re going to understand the process of commoditization,
then our understanding had better account for the difference between the market
for milk and the market for DVD players.
One way to understand that difference is to claim that all
goods and services either are commodities already or will eventually become
commodities, and that goods and services in markets with huge price
differentials simply haven’t become commoditized yet. This seems to be David’s
stance, though he doesn’t explicitly say so. The problem with this stance is
that some markets appear to never become commoditized, and there is no real way
to see how they can become so. Consumer electronics has resisted
commoditization since the invention of the radio, and there’s no sign that
commoditization is near. While overall prices in the market for consumer
electronics have dropped steadily over the years, a wide variety of prices
continues to be a feature of that market. Indeed, if anything, the price
differentials have increased, not decreased.
Another way to understand the difference is to note that
some goods or services lend themselves to something called “product
differentiation.” This concept is predicated on the idea that consumers have a
wide variety of tastes, and that some are willing to spend more than others to
satisfy those tastes. Goods and services marked by product differentiation
tend to compete on the basis of “features” rather than price. Economists tend
to not like this idea, because it makes for very messy mathematical models.
That David’s argument fails to grasp this distinction is
bolstered by the fact that David quotes Karl Marx. Marx, along with most of
his 19th century contemporaries, was mired in an effort to arrive at
some objective notion of “value,” and all of Das Kapital is predicated on the idea that this objective definition
exists. Without it, the whole idea that the proletariat even has something
that can be called an “excess value” that can be appropriated by capitalists fails
to make any sense (which is probably why so many people fail to get past
But I digress. Our question is whether the market for
computer software (which is a stratified market) more closely resembles the
market for milk or the market for DVD players. We might note the lack of
variety in, say, desktop software, and conclude that software more closely
resembles the market for milk, but I think that would be a mistake. The
difference between a commoditized market and a non-commoditized market does not
lie in the variety of products in the market at any given time. Rather, the
difference is whether or not products within that market lend themselves to
competition on the basis of features.
I don’t know that there’s any more conclusive evidence on
this issue than the fact that everything we talk about in the software industry
has to do with features. Every religious debate from vi vs. emacs to Macs vs.
PCs is really a debate between differing individual preferences. Even the
discussion involving “total cost of ownership” is really a discussion about
features, or, more precisely, it’s a discussion about the way in which certain
features justify a differential in price. We can “taste” the difference, which
means the market for software is not like the market for 2% milk.