Use the links after each section or scroll.
Layoffs of another 5,000 people and closure of one of Sun's biggest campuses in Newark Ca. were announced yesterday in reponse to pressure from shereholders and analysts for Sun (SUNW) to return to profitability. This brings the headcount to around 32,000 and the stock price to just under $5 from a peak employmnent in 2000 of 52,000 and a stock price peak of $127 which split to $67 almost at the same time.
I worked at Sun from 1997 until 2004 and experienced a persistant problem of disconnect between management and individual contributors that cuased me to move across a major org boundary in 2000. The same issue caused me to give up and leave by 2004.
In 2000 I remember a discussion I had with several people about overall stretegy and where we fit in with our individual experience and talents. The discussion was about how unique Sun's place was among its competators. At that time we felt that the microprocessor and operating system Sun used would keep it distinct from the competition. That Solaris and SPARC were superior enough to Intel and Linux and Windows to not pose an immediate threat to Sun in the server and high-end UNIX platform market. We also knew that there was a shift taking place from legacy technology to intrgration around the web and Java and that resources would be directed at these new opportunities. Time would prove that these assumptions would disadvantage Sun.
A question to ask is how unique is Sun now? I remember how Sun rose from obscurity by providing an inexpensive UNIX based (BSD) graphics workstation aimed at engineering and scientific users. Its advantage was the superior reliability and performance of its OS and compilers. Having a good fortran compier was and is reamins an important selling point. I am not sure what distinguishes Sun anymore.
TopAs it came to pass, Intel and related microprocessors beat SPARC much sooner than anyone at Sun predicted, and what is worse Sun made a near fatal mistep by trying to discontinue Solaris support for X86 in 2003, which it had to reinstate. SPARC had performance and reliability issues all throughout this period while Intel was scaling into the space once dominated by Sun servers. Sun tried to compete by offering Solaris and its Linux in low end blades and Solaris on its big servers offering an integrated setup.
TopSolaris had an advantage over Linux of scalability and security, which on the big corporate servers made up for any concerns about supporting every applications that might be available on other UNIX like systems. It could support thousands of users and stay up for years and be very secure provided the system administration staff was knowledgable, and there were many GUI based support tools to monitor and help people manage these systems. With a java runtime tuned to perform at its best on a platform Sun controlled, web services could be delivered on a large scale. It could still interoperate with linux on a Sun supplied blade making for very good performance and security integration end to end and also deal with third party clients running oher Linux and Windows OSs.
TopSun has really never made any money on Java. It gives it away hoping to make up revenue by selling a development environment and a web server integrating it. The problem was that the competiton was very much ahead of Sun in delivering these premium services and by the time, 2003, when Sun was poised to compete, it did not have the manpower to keep improving its offerings fast enought to out do its competitors, IBM, HP, etc. If the plan was to make the hardware and the operating system commodities, the plan failed miserably.
TopThere has always been a problem with message at Sun. First, it is due to complexity and confusion about what is offered. Rather than having a sharp focus and a clear message Sun tried to be everyting to everybody and it failed when they tried to be a system integrator at a time when layoffs were reducing the number of people who could deliver. Part of this is due to disconnect between the major sections of the company, the corporate strategests not listening to engineers and the latter being igmored by managers who were promising things to sales who were claiming reliability to customers that couldn't be delivered. Things might have been better had it been clear that Sun was about delivering Solaris on SPARC and Intel, but that became lost as Sun moved on to immature integration tools that were not as good as competators while it open sourced its reliable legacy offerings.
TopI suspect that the management problems I saw as an individual contributor at Sun are common to many companies of the same age and potential, but there are ways to avoid some of the pitfalls I eperienced.
I think the the best test of management happens when things go wrong and when financial circumstances force one to do with less. You see in who leaves and who stays how astute the managemnt is about the qualities of people it needs to keep. You see in how well the managemnt handles change where their priorities are.
The most persistant problem I had at Sun is that I didn't think that most managers appreciated how much effort it took for an individual to get and develop his or her individual skills. What I encountered at least twice is the view that you will change because we told you to with no regrad to what you have brought or could help someone else to do. That was left to you to find your fit. Under a hiring freeze and a culture of organizational complexity it is hard for an individual who was used to concentration and some isolation to branch out and find where he could help better somewhere else. I moved across the org chart from front-line support to sustaining engineering because I felt that the managers were jerking us around and had no respect for what we knew. Eventially, the feeling that line managers were just making it up as they went along with little regard to intellectual capital forced me to leave Sun.
In this sense, if a business facing hard times was to make due with less, heap more roles on each person, it would have made sense to pay more attention to skills match and rather than punish people for not changing when commanded to, to cherish valuable skills that might become more valuiable as business direction is forced to change, especially when a retrenchment to legacy skills is exactly would rescue the company from a failed strategy.
Finally, up until Ed Zander left in 2003 there was an anonymous back channel where individuals could point out problems to senior management without fear of reprisal from line managers. That disappeared with no fanfare that Summer. That was the sign to me that I would be leaving Sun. It was also the sign that Sun would self-destruct sheding talant is needed under management who didn't really understand where the value lay in the people they managed and would use the least relavant criteria to determine who left and who stayed, prehaps being concerned about who would cost the health insurer more or who was pushing retirement age, rather than who could really help the company.
TopSun should take a clue from Apple who seemed on the ropes a few years back but rebounded by going back to its core competancies. Under Steve Jobs with Mac OS X based on BSD UNIX Apple has done excactly what Sun could have done with Solaris on SPARC and Intel, and like Apple with the Power PC, Sun could have decided to drop SPARC for Intel, but garnered support by providing rock stable server platforms that did automatic upgrades and have almost no hardware related support problems because of control of the hardware platform. Instead Sun may go the way of SGI which has recently gone out of business or more like that of Pyramid Technologies who was bought out. Sun would fade into the sunset as the last of its support contracts need to be honored.
TopWhat are the lessons of this? There are several: 1) You need the trust and support of the talented and smart people who create and maintain your technology. 2) You need to cover your bases by keeping experience from your history and legacy so that if a new iniative fails you can retrench by restablishing yourself from a proven base. 3) Encourage internal criticism of stretegy, of your marketing and PR. 4) Encourage constructive skepticism. 5) Do not give top-down management too much power. 6) Provide back channels for concerns that do not depend on chain of command. 7) Do not allow any grievance proceedure to use the same chain of command that created the policy that is disputed. 8) Screen management candidates for personality disorders expecially narcissistic and antisocial disorder and do not allow these people to manage. 9) Do not allow bottom line or accounting concerns to peovide an excuse for tacit age and health care discrimination. 9) Make skills assessment and tracking an important part of personal and headcount management. Make it independant of the org chart. If an employee has a specfic skill and is available, even part-time, his skills should be known and used, especially in a period of headcount restriction. The personality or incompetance of a manager should not be an impediment to this.
Top