Greetings Economists,
On Feb 2, 2007, at 1:33 PM, Michael Perelman wrote:
that the company has a tightly
conceived model of what it wants. Separate teams assemble discrete
parts that are
expected to fit together like lego blocks. He says that they were
successful in
doing this.
Doyle;
If one looks at Cusamano's reports they are trying to deal with the
complaint of software re-use. The more tightly bound the software to
discrete projects the less likely it is used for other projects,
therefore the market gets constricted and quickly obsoleted. This is
summed up in the term 'reuse'.
quoting a blurb about Cusumano's analysis:
If you are starting out in the software business, Michael Cusumano has
some advice for you: think hard about whether to specialize in products
or services, and take a serious look at trends in the industry. For the
past 20 years, Cusumano has been consulting with and researching some
of the top high technology companies worldwide. He has seen software
companies enter the market with a “killer application,” make some good
profits, only to find their product “commoditized.” In one case he
describes, the price of a software license fell from $1.5 million in
2000 to $250k today. And, he cautions, “in bad economic times, product
sales can fall off a cliff.” So what works? “The only guaranteed
revenues for software companies may be services and maintenance
revenues,” Cusumano says. In fact, his research shows one company after
another (PeopleSoft, Oracle, SAP, for instance) transitioning from
products to services in order to survive. Cusumano has graphed so many
of these corporate “criss crosses” that he considers them “life-cycle
models.” Only one company, he says, has managed to stay exclusively
true to product sales: Microsoft. To succeed, “most software products
companies become services or hybrid companies…where you’ve got the
basic product and build some custom features or a special interface, so
the solution you’re selling becomes much stickier.”
Doyle;
The discrete parts idea is still motivating Simonyi in the intentional
software companies springing up. I think the complex dynamic of
general to specific is what is driving this process. Software like
what Microsoft makes is supposed to be general purpose, everything but
the kitchen sink, but a generalization of the process would not yield
software so inaccurate and tied so strongly to a rigid market
structure. The investment costs to change from XP to Vista is 6
billion. Or Intel being caught flat footed by AMD. Once committed to
a strategy so far everyone seems to be unable to avoid the statics
problem in the models. Microsoft junked Simonyi's ideas because it
would have gutted dot net. IBM going to services touches on how
flexible human workers are compared to the software and hardware.
Doyle
