Tuesday, 7 July 2009

Grieving Over the Death of a Cash Cow

A sizable portion of the software cash cow herd is dying and our industry is learning to cope with grief, the natural expected response to death. Grief is painful, but absolutely necessary and completely unavoidable. Shock, anger, guilt, obsession, depression, confusion, and feelings of helplessness, emptiness, and loneliness are among grief's many facets. It's almost too much to bear.

Once upon a time, vast herds of carefully controlled cash cows were tended with loving care until some of the cows mutated and escaped. Actually, let's be frank, the cows were deliberately genetically engineered and then set lose with the disruptive intent, e.g., to undermine competing herds of cash cows. Those with the biggest herds were often in the best position to sacrifice a few cows and could draw attention to the virtue of their sacrificial offerings. Consider though how much it cost to develop what's at Eclipse today, i.e., hundreds of millions. No doubt this is dwarfed by commercial investments in Linux, i.e., multiple billions. Do these appear to be purely noble acts intended to build a grand public edifice?

Most certainly open source investment is generally not a selfless act; yes, it really does involve investment and yes, there is virtue in it as well. However, whenever a valuable resource is consumed, a return is expected. It is the economic way of the things. There can be no exceptions for long. So we must wonder, who will continue to tend the free cows and what return on investment do they expect? I.e., what is the emerging business model for our future software economy? In other words, how will Eike stock his bar so the goodness of CDO continues to flow?

I've already suggested that open source can act as a way of targeting competitors by undermining their cash flow thereby making them less of a competitive threat. Of course that tends to commoditize the domain so that no one makes money from it. A related strategy is to open source only part of the solution, just the basic frameworks and capabilities, and then provide value added software to drive revenue, i.e., a free cow complemented by a cash cow. In principle this can be quite effective because consumers love free things. From that perspective, open source software is almost like a coupon. If it could be contained to that, we'd not need a new business model, so this approach is a pinnacle of success and has shining appeal.

Another approach is to just open source everything. After all, we all know that free software still takes effort to install, learn, use effectively, and keep up-to-date, so there's still money to be made in the services around all this "free" software. Unfortunately, all this free software has produced a debilitating almost viral condition known as freetardation: the irrational belief that paying a dollar for anything software related is offensive in principle. How many organizations and individuals don't end up spending person weeks of time and effort doing poorly something they could have accomplished more effectively with appropriate technology or expert assistance at a small fraction of the real cost? The old penny wise pound foolish principle often rears its ugly head particularly when all the costs and benefits are not properly measured. The problem with the service model is that it's not a license to print money the same ways as is selling software. Software is practically infinitely scalable, i.e., just provide enough download bandwidth, while service generally requires more people, highly skilled people who are themselves expensive and in short supply. Still, it's also a proven model that's held up well.

A valuable aspect of open source software that's more easily overlooked is as protection against vendor lock-in. A few weeks ago at one of the banks it was explained to me that they're fed up with the products they're using. They're too expensive and provide little value over and above what's available in open source. What they're most frustrated with is that each year the salesmen show up with a laundry list of cool new features that justify the upgrade price and yet doesn't include a single feature they actually need. Open source provides better cost control and a better hedge against arbitrary future changes. Self service is even possible and while open source software too can change, there's always the option to fork. Decision makers are in a much better position these days to control their destiny.

I'm not really sure how this will all unfold in the longer term. I doubt anyone is, but the trend toward open source is clear. Probably some of the cash cows will survive, perhaps the value-add approach will work well, at least for those most skilled at riding the leading edge of the crashing wave, but most certainly the service model will survive regardless, because all the models rely on service. This should give committers pause for a bit of self reflection.

Personally I'm not grieving, at least not over the demise of software cash cows. Like other committers at Eclipse, I am a highly skilled expert, and that puts me and professionals like me in the driver's seat. We fit into the economic picture regardless of what business model prevails. If you think of software as content, much like a song, book, or movie is content, then we are the content producers. If you think of software as a service, we are the service providers. There is no software without us, and no, ten unskilled workers will not produce the same result as will one highly skilled individual. So while employers will be tempted to farm out your job to someone else if they think they can save a dollar, they're definitely going to need someone just like you. By being the best at what you do, you're in an excellent position to control your destiny because those in need of expertise will often be willing to pay top dollar to get the very best.


Donald Smith said...

Amen. And thank you for introducing me to the term "freetardation", I shall use it mercilessly :)

Matthew Hall said...

Joel Spoelsky has this to say on the economics of open source.

Mark Coggins said...


I don't know if I can bear your post:




Robert Wloch said...

Ed, amen and thanks for your wise words as well as your word creation!

Madhu said...


Thanks for the nicely written article.

Recently I attended an exhibition of lot of startups in Bangalore (http://eclipse-info.blogspot.com/2009/05/startup-city.html).

One of the common pattern found across the new products offered are that most of them are offering the services for free to the common public. Their revenue generation plans varies from advertisements, charging for specialized services, etc..

If we take eclipse itself, our community is one of the best place to advertise the new business specific products (take xtext for example).

teeth whitening said...

The problem with the service model is that it's not a license to print money the same ways as is selling software.