The X Factor: You Can’t Steal What’s Free, But You Can Pay a Lot for Something That Isn’t Worth It
December 4, 2006 Timothy Prickett Morgan
Back in the early 1990s, when IBM was on the rocks after its mainframe business imploded and the world shifted to cheaper Unix servers and client/server development was just getting underway, the Wall Street Journal ran a multi-page story that I have never forgotten. The story was one of hubris, of Big Blue’s lack of understanding of how the market had changed so much around it that it took the company the better part of six years to find its feet again.
Had it not been for the Y2K and dot-com bubbles, IBM might not have found its feet again. That’s how bad it was, and it was only the bubble that allowed IBM–and all of its key competitors–to grow their sales and profits. In any event, it was no wonder the company brought in an outsider, Louis Gerstner, to run the show from Nabisco. Gerstner is now chairman of defense contractor and private equity firm The Carlyle Group, and he held Big Blue together with baling wire and his own personal strength as he sorted out the company and figured out how to retrofit it for the modern era. That Gerstner saw that IBM was not, as many had said, worth more separately than it was worth together is obvious, since he did not break the company up as many had been advocating at the time. It is hard to do otherwise but conclude that Gerstner saw that IBM was such a mess that if you broke it apart, it would be worth less than the sum of its parts. Yes, it was that bad.
The customers knew it, too. They were on the front lines, and they were looking at all of the alternatives as IBM had launched its ES/9000 mainframes and was trying to charge $100,000 per MIPS for them. In that Journal story, there were two immortal lines from customers that I have never forgotten because they were so funny–and so true. The first was from a prominent IT manager, who asked to remain anonymous. To paraphrase the first one: “No one ever got fired for buying IBM. Annnnggghhhh [making the sound of the big red X on the game show Family Feud.] Next contestant, please.” The second one was even funnier: “IBM–you can find better, but you can’t pay more.”
To fix itself, IBM became a services company–at least as far as its bookkeeping is concerned. But don’t be fooled. A lot of its sales and profits come directly and indirectly from sales of its four major platforms–mainframes, System i and System p Power-based servers, and System x X64-based servers.
Survey the elements of the modern corporate desktop and data center software stack. So much of the software is still firmly in one vendor’s control, even after five or 10 years of open source alternatives, which charge for support but not for software. But every year, the appeal of open source gets stronger as the advocates of those older technologies move on. Some day, Unix will seem proprietary–and anyone looking at the price tags of anything except Solaris 10 probably believes that now. Add in the effects of Software as a Service and, eventually, utility computing, both of which will drive up efficiencies and wring a lot more costs out of IT infrastructures. The companies fighting to bring such technologies to market can probably make money hosting proprietary software made by someone else. But they will make a lot more money if they used a mix of hardened open source tools.
In the long run–maybe a decade from now, but certainly not this year–the battle lines will be more clear, and companies will look at the capital and operating expenses associated with their data centers and desktops and the applications they run on, and they will start looking at alternatives that bring the costs down. They have no choice. The lessons of the dot-com boom have taught CEOs that CIOs had a lot of fat, and now they are being made to run it off. Every year.
You get an inkling of this shift when Larry Ellison, the chairman of Oracle, which just spent a fortune amassing a huge portfolio of application software stacks–PeopleSoft, J.D. Edwards, Siebel Systems, Reitek, and on and on and on. Buying up customer bases and getting the maintenance money is a smart move in the 1980s or maybe even the 1990s. But then along comes something like Salesforce.com for customer relationship management and Workday for what will eventually be a full ERP suite–available through the SaaS distribution model. (Workday was launched two weeks ago by PeopleSoft founder Dave Duffield, who is still smarting after the Oracle hostile takeover almost two years ago.)
While Salesforce.com and Workday, among others, are getting IT used to the idea of SaaS, it won’t be that hard to make them consider true utility computing. And by then, it is not hard to reckon that someone would have been smart enough to create an open source ERP suite that can be deployed with customizations in a virtualized machine using open source and open standards. And at that moment, what is all that software that Oracle has paid a fortune for worth? If all that Oracle wants is a bigger piece of the pie now so it can figure out what to do later–which is not a bad strategy if you can afford it–then the company will get exactly the results it expects. It will have bought some customer bases and some time.
Oracle, of course, is not the only company feeling this pressure. So is Microsoft, which is looking at an installed PC base that doesn’t want Vista the same way it wanted Windows 3.1 or Windows NT Workstation or Windows XP. And when Windows Server 2003 Small Business Server R2 costs $599 with five seats at street price compared to Novell‘s SUSE Linux Enterprise Server 10, which costs $249 for an unlimited number of seats, how long before small businesses are going to start playing with Linux? When they hire their next IT personnel, they will know Linux as well as they know Windows.
Which brings me back all the way to Unbreakable Linux, which shows a tactic that vendors desperate to keep their slice of the IT budget will be able to do. With so many commercial-grade open source projects out there, it will be sorely tempting to take the stack and offer support for half or less of the cost of the official commercial version of the product. This is essentially what Oracle is doing with the Red Hat Enterprise Linux stack, which it is offering as Unbreakable Linux under Oracle’s own aggressively priced maintenance and support terms. Oracle could also offer Unbreakable SQL, based on MySQL, and Unbreakable JBoss, based on Red Hat’s JBoss Application Server. Why stop at the operating system and the applications that come in the distribution?
The real question is whether or not support for open or closed source programs–desktop or server, large or small–are really worth the money. It takes people to make and support software, and open source has certainly undermined the revenue base on which companies can depend to cover the cost of development and support teams. And cheaper open source alternatives–which rely on the kindness of the meritocracy behind a project to do a lot for free–puts pricing pressure on all kinds of software. The cost of having human beings who understand increasingly complex programs and their interactions–who are at our beck and call–is high, even with offshoring and open source.
Maybe SaaS and utility computing will be enough to make enterprise software cheap enough to make the CEOs and bean counters happy. (If pay-per-use is attractive for software, it is compelling for hardware, particularly if it is underutilized as it is in many organizations.) Maybe even that software will have to be open source to drive costs down further. But in the end, is the ultimate goal to have software that costs nothing, or a set of business processes that add value to the organization.
The title of this essay was a smart-alec way to remind people to think about the value of the software they buy. Oracle can’t steal Red Hat’s Linux software business–at least not entirely without killing it or acquiring it. But, if Oracle can come up with adequate or superior support for Red Hat’s own products, it surely can make some money and put its technical people–who know a lot about Linux–to better and perhaps more profitable work while also leveraging a new product in its vast installed base. But, someone can turn the tables on any proprietary software player and offer SaaS or utility computing services, buying products in huge volumes–driving down a vendor’s sales–and making money on the spread. And someone who does this trick using inherently less expensive open source technologies–since these companies spend next to nothing on marketing–can really hurt the established software players.
It will be interesting to see how this all plays out.