April 8, 2003

There may never be another software billionaire

- By Robin 'Roblimo' Miller -

We often forget that that idea of software as a boxed, mass-market, proprietary product is only about 30 years old. During these three decades, the software industry has made a number of individuals wealthy. But this era may be over. The future of software may not have room for new billionaires, even though there is a growing demand for new software, and opportunities for people who write that software will continue to grow along with that demand.

Profits are getting wrung out of the software business. In story after story, we hear about Linux taking over a growing percentage of the server operating system market while only generating a tiny percentage of operating system sales revenues.

Sun can fume, SCO can sue, and Microsoft can attend all the Linux conferences it wants, but these are nothing but holding actions. Linux is a "good enough" server solution for the vast majority of corporate users, so its server market share will continue to increase. Users aren't concerned about proprietary operating system vendors' problems. They want the best value they can get, and that is almost always Linux, with *BSD as a ready alternative in areas where Linux falls short.

What about Red Hat's Enterprise Server?

Red Hat's Enterprise Server will never make Red Hat as much money as Windows Server and IIS have made for Microsoft, because it offers an alternative to "do it yourself" server building and maintenance, not a set of proprietary products. You buy Red Hat's top-of-the-line product because you believe it is cheaper than rolling your own, not because you are locked into it.

And if you want enterprise-level Linux in a box, but don't want to use Red Hat's version, SuSE, Mandrake, and many others will happily accommodate you. This competition holds prices down -- and forces Sun, Microsoft, and other proprietary operating system vendors to either cut their prices or give up on the server operating system business. Either way, this segment of the software industry is likely to be much less profitable in the future than it has been in the past. The new opportunities in this area are likely to be in service and support, not product sales, and these are low-margin activities with low entry barriers that allow small companies to compete evenly with large ones. Indeed, small companies often have an advantage in service businesses because they are not saddled with big company overhead or management inertia.

Upshot: Stop looking for billions of dollars in the server operating system market. Instead, look for small, service-based opportunities that may make millionaires out of some smart entrepreneurs, and plenty of decent salaries for competent software developers, sysadmins, and support technicians, but don't expect a single company or small group of companies to dominate this area.

Desktop/client changes

Let's start by saying, right up front, that the current Linux desktop environment is not as slick as Windows or Mac in many ways. Linux software is much harder to install for most people than Windows or Mac software, and there is a lot less of it available. Simple functions like changing mouse settings and downloading and installing attractive screensavers are needlessly complex. Requests for help in performing these tasks in online Linux forums are often met with screeds about the superiority of command line operations over GUIs or rants about why everyone should use Debian or Gentoo, rather than with useful advice.

Still, a home or small office user with simple computing needs can save a great deal of money using Linux, and can have a faster, more stable desktop on cheaper hardware than he or she can get with Windows or Mac. And in corporate or government offices where desktops only need to perform a few specific functions -- usually by accessing custom applications on a remote server -- Linux is an obvious choice for any cost-conscious manager.

Free OpenOffice and low-cost StarOffice may not offer as many features as Microsoft Office, but you don't see police officers driving around in Lincoln Town Cars or Cadillac DeVilles. They are issued Fords and Chevrolets, which cost a lot less to buy and maintain but do everything a cop needs a car to do -- in a professional sense, at least.

Intelligent corporate and government managers -- and millions of individuals -- are starting to think more pragmatically about their desktop computing needs than they have in the past, and the number that realizes they can do everything they need to do with free and open source software increases every year.

This doesn't mean Microsoft or Apple is going to go away, but that Microsoft is likely to come out with at least a few stripped-down, low-cost products, and that one day we may see Mac OS X available for low-cost I86x hardware.

We're seeing downward price pressure in the desktop software market at both the operating system and applications levels. This price pressure is going to become more severe in the future, not only because of Linux and open source applications that run on it, but also because proprietary software companies like Corel are hungry to win Windows customers and are willing to cut prices to do so, as witnessed by the increasing percentage of "major brand" Windows PCs that now come with Corel WordPerfect instead of Microsoft Works in their basic home/office software package.

Once again, we're seeing profits wrung out of the software business, which has been tremendously lucrative because profits margins for proprietary software have often been 75%, 80% or even higher.

Lower margins aren't going to kill the proprietary software business, but they are going to make it hard for software to create many new billionaires. This will not stop new software from being created. We see plenty of consumer electronics innovations even though margins in consumer electronics are thin, and we see plenty of new restaurants springing up even though the odds of success in the restaurant business aren't very high.

The software business vs. the restaurant business

If you're not traveling, nothing stops you from preparing your own food instead of eating out. And even if you're 2,000 miles away from your kitchen, you can live quite acceptably on cereal and sandwiches and other store-bought foods instead of going to restaurants.

The next bump up is fast food. You could easily say the fast food industry has wrung profits out of the restaurant business by making it harder for traditional sitdown eateries to attract customers -- and charge them enough to earn a profit by serving them. You'd be right. The percentage of "real" restaurants has dropped radically (at least in the U.S.) over the last three or four decades, while it seems fast food places have popped up on nearly every corner.

And yet, the best and best-run sitdown restaurants still manage to survive and profit despite competition from low-cost fast food stands and from eat-at-home, self-prepared food. And new ones open, and some of them make a decent living for their owners, and some of them even grow into chains or franchises that make their founders into millionaires.

The food business is essentially open source. Everyone -- including cook-at-homes -- has access to the same ingredients. Recipes are easy to create -- and to reverse-engineer. I can make exactly the same prime rib meal at home I recently paid $27 for at Lawry's. Lawry's even publishes some of their most-requested recipes on the Internet to make it easy for me. But still, there I was last week, sitting happily in their Art Deco Las Vegas restaurant, enjoying the ambiance and service -- the parts of the experience I would have trouble duplicating in my own kitchen -- and happily paying for the privilege.

In other words, I was paying for service and support.

Indeed, I was paying a hefty premium for Lawry's service and support. I could have paid a lot less for lesser support elsewhere, as in going to the restaurant in the hotel where I was staying and ordering their $15.95 prime rib special that, like Lawry's "signature" meal, included a salad, potato, and Yorkshire pudding.

Eventually, I expect the software industry to become more like the restaurant industry, with free software developers as analogs of at-home cooks sharing recipes, low-cost generic software packages (like StarOffice, WordPerfect or packaged Linux distributions) acting like fast food, and a few proprietary software producers (Oracle, Microsoft, Adobe et al.) maintaining a small but noticeable position at the top end of the marketplace.

Another restaurant parallel is the software industry's increasing dependence on low-cost immigrant (or overseas) labor instead of depending on the traditional -- expensive -- U.S. work force. The restaurant industry made this shift long ago, as did agriculture before it. This is going to be hard on U.S. programmers, but not many of them complained when restaurants started hiring (often illegal) immigrants for many kitchen positions or refused to buy imported cars when they started flooding the U.S. market, so politicians are not going to listen very hard to their complaints about being replaced by low-cost foreign workers. U.S. programmers are going to have to learn to live with the same job problems U.S. workers in other industries have faced for many decades, and since the vast majority of U.S. software developers still seem to dislike unions and claim to love "the free market," there is little chance of this trend reversing in the near future.

There will still be star programmers who make big buck$

You don't need to be a great chef to work in the kitchen at McDonald's -- or Lawry's. You just need to be a cook who has basic skills and can follow instructions accurately. It takes more skill and training to cook at Lawry's, so pay there is better and turnover is lower. But there are restaurants that rely on their chefs' reputations to command high prices and fill their dining rooms, and even McDonald's needs someone at the top of the food chain (sorry; couldn't resist that one) who is responsible for creating recipes instead of merely following them.

Real chefs often command high -- even huge -- salaries. A few stars even get (usually cable) TV shows, book contracts, and their names on "signature" kitchen accessories the way basketball stars have their own shoes, and become multimillionaires as a result of their stardom.

But most culinary school graduates never become star chefs. They live in modest circumstances and drive used cars -- and work extremely long hours in return for their salaries, and have less job stability than they'd have in almost eny other occupation besides construction work. Or computer programming.

Still, classes at the culinary schools are full of eager students wearing white chefs' uniforms, just as computer science and programming classes are full of eager students wearing "got root?" t-shirts. Most of these students know their chances of becoming superstars in their chosen field are as slim as the chance of a randomly-selected college basketball player making it into the NBA and getting a lucrative endorsement deal from Nike.

Cooking, programming, and basketball are all fields you must pursue as much out of love for the endeavor itself as for money if you are going to be successful in them. They are all hobbies for many, too, who do them purely out of love while earning a living doing something else. To become a star, you are not only competing with people who have professional aspirations, but also with talented amateurs, so you need to not just be good, but damn good, to rise to the top.

No one has a "right" to become a star basketball player or chef, no matter how much training they get. No one has a "right" to be a star programmer, either. Most programmers will always be the software equivalants of cooks, not chefs, following orders and doing routine work.

Too many programmers won't spoil the code

As computers become more ubiquitous and as the IT industry becomes more important to the world than it already is, we're going to need more programmers because we are going to keep needing new programs. Some of those programs -- probably an increasing percentage -- will be written under free or open source licenses either by talented amateurs or by professionals at end user sites or by professionals who work for companies like Red Hat and IBM that make money from selling software service and support, not necessarily from selling the software itself. But there will still be plenty of commercial (proprietary) software written. In absolute numbers the amount will likely go up, not down, because the total world software count will probably increase faster than the rate of change to open source if current trends prevail.

In other words, Microsoft and other proprietary software vendors could lose market share but still increase their absolute number of unit sales.

The big question is, "Will proprietary software vendors be able to hold onto their high profit margins?"

The answer to this one is, almost certainly, "No."

Today we routinely see news of $1 million proprietary Unix hardware/software server installations being replaced by $300,000 or $500,000 worth of generic hardware running Linux. We're also starting to see reports of $1 million, high-end proprietary desktop software installations being replaced by $200,000 or $100,000 in Linux and open source (or low-cost proprietary) applications, and we can expect this trend to accelerate over the next few years.

The end result is that a lot less money is going to be spent per installed program and per line of code than is being spent today. This does not spell doom for programmers, although it's possibly that most of the rejoicing among programmers over new work opportunities is going to be in India and China rather than in North America and Europe. But the real shakeout is going to come at the tops of companies as investors who are used to seeing huge software markups (and huge software profits) ask software company CEOs to perform a literally impossible task by fighting industry trends that are beyond their control -- and fire managers who fail to do the impossible, which will be nearly all of them sooner or later.

Oh, a few will survive, and some will even prosper in the new software business environment, just as Outback Steakhouse and a few other restaurant chains have managed to do well despite the long odds against success in the "sit down" restaurant business.

But these survivors are going to have profit margins and financial statements that look more like Outback's than like the Silicon Valley-type billionaire high-fliers we have associated with the software business in the past.

In other words, the software business will still be with us. As a business. A normal business, like any other, where efficient operators have a fighting chance at success and all others fail.

Whether this is good or bad depends on where you sit in what is increasingly being called "the software ecosystem." The only sure winners are software users, who will get more and better software per dollar spent. On the vendor side, there is going to be plenty of anguish, all the way from the tops of today's largest software vendors down to new computer science graduates looking for their first jobs, who are already dealing with some of the business realignment pain others in the software industry have not yet felt, but can expect to feel before long.

Click Here!