September 30, 2003

What Sun needs to do now to survive

- by Chris Preimesberger -

Yesterday's news about Sun Microsystems' huge net loss of $1.04 billion in its fiscal Q4 -- and more of the same expected for this current quarter -- puts squarely on the table some serious business issues that the company needs to address now if it intends to survive.

Even though the Santa Clara, Calif.-based company tells us that it can do everything it wants to do in hardware, software, networking, embedded systems, wireless, and now the desktop, its stock price -- which has been hovering in the low single-digits for nearly two years ($3.46 at Monday's close) -- ought to be a clear sign to management that it seriously needs to reassess everything it does.

Now in its twenty-first year of business, Sun long has been one of the most innovative IT companies in the world. As it became extraordinarily successful in the 1980s with its Unix-based workstations, and in the '90s with its Java-related software and licensing, it acquired other companies and expanded into other product areas. And, having taken a recent tour of Sun Labs in Mountain View, Calif., I can attest firsthand that Sun has many ingenious new ideas in early and mid-development.

However, even with all its past success and dozens of new ideas on the horizon, it is clear now that the company is juggling far too many balls at once. Something is going to drop; and lately it's been small packets of employees here and there. The casualties will increase unless some serious changes are implemented. Investors won't be pleased with continuing P&L statements like this last one.

Four major problem areas

There are four major high-level problems Sun now faces and needs to solve. I'll go out on a limb and say that if the company doesn't solve these issues within three years -- or perhaps sooner -- we'll be seeing Sun's Chapter 11 coverage here on NewsForge.

1. Too many products. Sun wants to do everything under itself, pardon the pun. It needs to pare down what it does, and then do those things better than anyone else. Is it high-end Unix workstations? Maybe. Is it Java servers and related software? Perhaps. Is it wireless and embedded software and processors? Too early to tell. Is it all of the above? Probably not. It's obvious the company is spread far too thin.

2. Marketing. Too many different -- and often conflicting -- marketing and sales messages. Too many rebrands, causing products to lose their identities. A number of smaller companies were swallowed, and previously successful product brands morphed into a big Sun lump, where, again, they lost their identities and customer following. Sun should also stop hammering Microsoft -- it may deserve it, but it's way old.

3. Open source ambivalence. The company is still dancing around whether to incorporate open source (and therefore get the community involved and trusting it) in its product offerings. Sun has been divided for years on this one. I hear complaints about Sun's licensing policies every day. It needs to loosen up here, big time.

4. Turnover. Loss of key business and creative talent in recent months (namely former president Ed Zander and chief scientist Bill Joy). If they ever lose Dr. James Gosling, head of the group that created Java in the early '90s, it'll be curtains.

Ideas for Sun to consider

I've been following Sun's fortunes since 1995, when I worked as a contractor for the company as editor of an international sales-channel newsletter called "Sun's Hottest." I also served as Java editor at DevX before joining OSDN, and have done analysis on several developers' surveys about Java and other technologies. I see a few steps Sun could take now to help propel its business forward.

Select perhaps half (or fewer) of the products you now feature and put your best resources into them. Discontinue or sell off all the rest. Yes, it will mean massive layoffs. But talented people will survive elsewhere.

Don't say you're going to put all those various brands under one umbrella, start an initiative called SunONE, change it to Sun Services on Demand, then to Java Enterprise System -- all within a span of 18 months. Knock off the name-changing on products. When you introduce a new product, give it a name, then leave it alone.

Don't be afraid to say something has been discontinued. For example, when the iPlanet acquisition turned out to be a mistake, the company never admitted it and just let the name and brand of many of the downplayed servers kind of twist in the wind. iPlanet still has several products in the ponderous catalog now called Java Enterprise System, but they're taking a back seat in overall company marketing.

Try something radical: Release as much of the Java code base to the open source community as possible -- as soon as possible. Look what IBM is doing with Eclipse; Sun is far behind on that score. Of course, your licensees will be upset; work out some kind of long-term, value-added agreements with them.

Make sure you retain the truly creative people at any cost. Losing Bill Joy several weeks ago was not a good move. He was the most visionary person in the company. Better give Dr. Gosling a raise.

Click Here!