The commoditization of software

COMMENTARY--For years, hardware companies struggled under pressurefrom the forces of commoditization, as widespreadpopularity of standardized operating systems forcedthem into withering competition, the end result ofwhich was a phenomenal reduction in hardware prices. Today, the shoe is on the other foot, as software is commoditized by the appearance of free (as in cost) open-source software. Constructed through the joint effort of thousands of programmers around the world, its low cost serves as a strong challenge to proprietary software companies' business models.

Commoditization is a powerful force for consumerbenefit. In hardware, however, there was never anyrisk of free hardware pulling the economic rug fromunder the industry. Commoditization, taken to thepoint where software is practically free, would runcontrary to the interests of software developers whohave an interest in earning a living from their craft. In short, a market where the price of softwareplummets too low will rapidly drain the value ofprogrammers, and hence cost many programmers theirjobs and drain the industry of investment as investorschase more lucrative places to put their money.

The "de facto" free nature of open source
The last time I poked my nose into the open sourcedebate, a lot of people pointed out that open sourcedoes NOT imply no cost. They noted that licenses canbe written in such a way as to require fees which arepaid back to the copyright holder. This theoretical possibility, however, does not change the fact that the vast majority of open source software is, in fact, available at no cost.

There are a number of likely reasons for this. Whenever you have access to the source code, cloningis facilitated. AT&T once owned the Unix operatingsystem, yet that didn't stop the rapid propagation ofUnix clones once AT&T showed willingness to licensethe full source to universities. Programming ideasare universal, and there are an unlimited number ofways to represent an idea in code. Thus, with careful attention to detail, developers can clone an entire program without directly copying a single line of code. Such cloning makes it hard to maintain a baseline price, particularly when certain segments of the development community dance to the tune of the free (as in cost) software movement.

Second, open source culture is suffused with a beliefin the merits of free, as in cost, software. As Inoted last June, RichardStallman is on record as saying that programmersneeded their expectations of compensation readjustedin order to return, Rousseau-like, to some lost statewherein programming Van Goghs do their work for thesheer love of programming. At the time, someresponded that I shouldn't pay any attention to theman behind the red curtain as he isn't representativeof the open source community at large. Yet,commoditization driven by free software is a growingreality, and every day freer, open source softwarebecomes available on the internet.

It's easy to disagree with specific statements made by Stallman. Yet, it is hard to avoid the fact that these statements, in the aggregate, have shaped the open source culture in such a way to create an emphasis on free, as in cost, software.

The nature of valuation
Value, or the price of a product, is not an inherentattribute of that product. It is the result of theinterplay between supply and demand, and is completelydivorced from any inherent characteristics of theproduct in question. Oil costs much the same toextract whether the price is elevated by war fears orlow from oversupply. Similarly, the fact that steelprices go up because of a rapid increase in the demandfor steel doesn't change much the raw cost ofproducing that steel.

Expectation, however, also plays a large part in the determination of value. If people expect to pay a high price, they tend not to question it. Any American who has seen the price of Levi Strauss jeans in Europe ($90-$100) has wondered in quieter moments whether they couldn't finance a future trip with a suitcase full of jeans at an entrance to Rome's Termini station. In Switzerland, the price of mixed drinks is astronomical ($10 is typical). A Swiss friend told me that this was a historical anomaly due to former high taxes on hard liquor in Switzerland. The tax today is more reasonable, but the price ofmixed drinks in bars remained because the Swiss wereused to paying those prices, and so bars were happy tocontinue charging them.

The same expectation principle applies in the marketfor software. Tales abound of companies that movedfrom a revenue model based on advertisements (hence,free access) to one based on end-user fees, only tofind that no one was interested in their productanymore. Adobe Photoshop is very expensive (currently$586.99 on Amazon), which is about $300 less thanit used to cost) yet it continues to be thebest-selling graphics editor on the market. EvenAdobe competitors benefit from the price. PaintshopPro is a relative bargain at $109, but even that priceis higher than the software average and is onlysustainable due to expectations cultivated by Adobe.

Living in a commodity market
If expectations are created that software isn'tsomething upon which consumers, business or otherwise,should spend much money, then that lowers the value of software, which by extension lowers the value of the programmers who make it. This isn't exclusive to makers of shrink-wrapped software, but applies up and down the supply chain, including even makers of tailor-made software (read: contractors and consultants). If you bought a coat for a buck, would you consider paying $20 to have it let out after too many fried chimichangas dipped in Cheez Whiz addeda few inches to your girth, or would you walk out andbuy a new one? On the other hand, if you paid $200for the coat, you might consider paying $20 to have ittaken out.

Consider SAP. Major consulting companies earnedbillions helping large companies customize the SAPproduct. SAP skills were at the top of the list onany jobs board during the 1990s. SAP was alsoeye-popping expensive, which justified customization expenditures that often ran into multiple millions.

The operating theory among many open sourceprogrammers is that the market for their services willpersist because people will want to pay forcustomization, something that will be easier to affordgiven cost savings in the base product. I'll grantthat customizing software isn't as simple as buying awider coat, but I reject the notion that companieswill spend MORE on customization (or even maintainformer budgets) simply because they paid less for thebase cost of software. More likely, expectations ofsoftware cost will be adjusted downwards, resulting indownward pressure on developer salaries (which ishappening) and a greater incentive to subletcustomization to the lowest-cost location.

The market for development outsourced to foreignlocations is getting a lot of press these days. Certainly the slow economy has a lot to do with thedemand. Furthermore, programming skills are moregenerally available than in the past, which was boundto happen given that nations such as India and Chinaaren't standing still. However, I think the role thatsoftware commoditization and its driving force, opensource software, is given too little attention. Alarge component of America's economy is informationtechnology, and free software undermines demand forsuch products, thus hampering recovery and increasingthe attractiveness of outsourced development. Furthermore, given the general lowering of softwareprice expectations initiated by the popularity of free alternatives, interest in outsourced development only rises.

Lastly, high profits attract high levels ofinvestment. RedHat managed a thinprofit selling open source software in a fastgrowing market for Linux, but how does that stack upagainst Adobe, IBM's software division, Oracle, oreven Microsoft? That sort of profitability attractsinvestment by the bucketful, whereas RedHat's sortwon't attract the same levels. Less investment meansfewer software jobs, and lower valuation of programmerskills.

Conclusion
As I've noted in the past, software is astrange market. Doctors don't roam the countrysideoffering free medical services and spewing invectiveagainst the evils of the for-profit medical industry. Carpenters aren't forced to contend with a free-carpentry movement, and financial analysts at MerrillLynch aren't pilloried at "free financial analysis" conventions. For whatever reason, a surprising number of people whose aim is to make a living from programming endeavor to provide free alternatives to product sold by others within the industry.

Free software has the same effect free TVs would havein the market for televisions. The question on theminds of people involved in the production of software(or TVs) should be whether that effect is desirable. As investmentin software shrinks, so will the value ofdevelopers. That's a reality that even open sourcedevelopers must face.

John Carroll is a software engineer living in Ireland. He specializes in the design and development of distributed systems using Java and .Net. He is also the founder of Turtleneck Software. We wish to give John a special thanks for his 25th contribution to our forums.