The headline on the front page of this morning’s Merc was about airline nickel-and-diming, but the lead headline on the business section was
‘bold’ and ‘risky’
Yes, Sun overpaid in blowing $1 billion on an open source database company that only generates $50 million/year in revenues. Perhaps I’d call that decision ‘bold and risky,’ but that’s old news.
What about open sourcing the rest of their code, including Solaris and Java? Sure, it would have been bold in 1998 or even 2001, but in 2007 or 2008? Not hardly.
By the time (2003) when we were doing our paper on Linux adoption, it was clear that Unix was being commoditized, and that cost-conscious buyers were increasingly unwilling to pay a premium for Sun’s products. Sun reminds me (to a lesser degree) of Digital Equipment, who came to us in March 2004 for advice on open sourcing OpenVMS (which they never did). As with DEC, Sun’s potential to capture the imagination of the industry (particularly software companies and major IT buyers) through open source has long since passed as every big company now has some sort of open source strategy.
The opportunity was when Sun led the industry. I remember back around 1996, I asked one of the IT guys at UCI whether he’d recommend FreeBSD or Linux. Ben said that FreeBSD had better technology, even if Linux got better press. While FreeBSD was good, Solaris was the industrial strength BSD variant that Unix sysadmins preferred. Even through 2000, Solaris was known to be superior to Linux by key IT buyers, so there was still a window to prmote an open source Intel Solaris using a paid support model, ala Red Hat.
Instead, Sun — like other established firms with established software revenues — worried about cannibalization more than growth of adoption, and watched open source rivals grow their share while Sun’s paid products occupied an every smaller niche. Sun worried too much and too long, given (as the Merc notes) it makes two-thirds of its revenues from hardware — and that it also makes significant revenues from services. As IBM has shown, both hardware and services revenues are helped by giving away software, and are helped even more by establishing your software as an industry standard through open source adoption
Is the headline right about Sun’s current strategy being ‘risky’? Sure it is, but not doing anything is certainly riskier. Sun once led the Unix and network computing industries, and has been unable to adjust to its post-bubble drift towards irrelevance. If it doesn’t do something to grab mindshare and marketshare, it will be like Cray and SGI and Data General (and eventually DEC), occupying a ever-smaller niche.
I still think Jonathan Schwartz has a plausible strategy. The problem is, the strategy isn’t working. After running up to $25 last fall, the stock is now near $13 — below where it was three years ago. Its shares fell 23% earlier this month the day after it announced an unexpected loss — missing expectations by 22¢/share.