Brandon's Blog

Sunday, September 17, 2006

"In the world of business, success is about who you know."

Saturday, September 16, 2006

"The Next Bill Gates"

http://www.businessweek.com/technology/content/nov2005/tc20051104_013385.htm

(could be a fusion of all singularity technologies, genetics (biotechnology), nanotechnology, and robotics)

Principles

"Acquire Research Results by Hiring and Buying
Conventional (loser) economic wisdom holds that monopolies should spend heavily on research because they are in a position to capture the fruits of the research. But if you want to become as rich as Bill Gates, you have to remember that it is cheaper to wait for a small company to come up with something good and then buy them. In the old days, antitrust laws kept monopolies from buying potential competitors. But not anymore. When Microsoft products were threatened by network computers and Web-based applications, they simply bought WebTV and Hotmail.
Another good strategy is to hire the right people. Some of the guys who wrote Microsoft Windows had previous worked on window systems at Xerox PARC. So Xerox paid for the research; Microsoft paid only for development.

In the long run a tech company without research probably can't sustain its market leadership. So you'll eventually need to build something like research.microsoft.com (check out netscan.research.microsoft.com to see some interesting online community research).

Let Other People Do the Programming
If you're a great engineer, it can be frustrating to rely on other people to translate your ideas into reality. However, keep in mind that the entire Indian subcontinent is learning Java, and that if Microsoft, Oracle, SAP, and Sun products simply worked and worked simply, half of the world's current IT workers would be out of a job. You're not going to get rich being "just a coder." Especially working in painful low-level imperative languages such as C or Java. It might be worth writing your own SQL queries and HTML pages since these tend to be compact and easier than precisely specifying the work for another person to do. But basically you need to get good at thinking about whether a piece of software is doing something useful for the adopting organization and end-user. Bill Gates does code reviews, not coding.

Train your new CEO
If you're an intelligent curious person it can be painful to run a company of more than 50 people. You spend more time than you'd like repeating yourself, sitting in boring meetings, skimming over long legal documents in which you know there are errors but aren't sure how serious, etc. The temptation is to hand over the reins to the first "professional manager" who comes along. And that's what the standard venture capitalist formula dictates. But Bill Gates didn't do that. He hired Steve Ballmer in 1980 and gave him the CEO job 20 years later. Making money in the software products business requires domain expertise and a commitment to solving problems within that domain. Great tech companies are seldom built by non-technical management or professional managers who aren't committed to anything more than their paycheck. Adobe is another good example. The two founders were PhD computer science researchers from Xerox PARC who were passionate about solving problems in the publishing and graphics world. They are still guiding operations at Adobe.
Note that this is a principle that Old Economy companies have long understood. Jack Welch joined GE in 1961 and became CEO 20 years later. Sometimes an Old Economy company may pull in a few outsiders to senior positions but, because they have such stable bureaucracies underneath, they can more easily afford this than startups.

See Charles Ferguson's High Stakes, No Prisoners (1999) for a longer explanation of how hired-gun CEOs manage to kill software products companies.

Focus on Profit
"At Hewlett-Packard, people, materials, facilities, money, and time are the resources available to us for conducting our business. By applying our skills, we turn these resources into useful products and services. If we do a good job, customers pay us more for our products than the sum of our costs in producing and distributing them. This difference, our profit, represents the value we add to the resources we utilize."
-- David Packard in The HP Way
Remembering to make a profit was tough in the dotcom 1990s but it turns out that Hewlett and Packard's ideas were right. Most of the management teams at dotcom businesses, by being disorganized, unintelligent, and ignorant, were subtracting value from the resources that they controlled.
How does one make money in the software products business? Simple. The necessary step is to build something that becomes part of information systems that generate value for organizations and end-users. Once you've created value you can extract a portion in lots of ways. You can be closed-source and charge a license fee. You can be open-source and charge for training, service, support, and extensions. But if you aren't getting your software product into important information systems, you don't have a prayer, no matter how slick your marketing materials.

If you're creative and diligent the software products business is extremely lucrative.

Let the Venture Capitalists Schmooze Wall Street ...
... but don't let them run your company. A profitable Microsoft Corporation brought in venture capitalists (VCs) at the last minute. They didn't need or spend the money but used the VCs to boost their valuation at the initial public offering, thus getting more money for the shares that they sold. Venture capitalists are dangerous because even the most successful might not know anything about business. Remember that there are tens of thousands of venture capitalists in this world. Assuming that they make random choices of companies in which to invest there will be a Gaussian curve of performance. Some firms will do consistently better than average even if everyone is guessing. Imagine that thousands of monkeys are flipping coins; some of the monkeys will get 10 heads in a row. These are the monkeys that will be celebrated for their insight. These are the monkeys whose track records will lead to uncritical cheerleading by underwriters and public investors. In bull markets such as we had in the 1990s nearly all the monkeys will be fairly consistent winners. But remember your next-door neighbor who made money in the stock market in 1985. He convinced himself that he had special insight and ability when actually he was only holding high-beta stocks in a rising market. So his foray into the commodities futures market wiped him out in the crash of '87.
Bottom line: successful software products companies spend most of their time listening to their customers and users rather than to venture capitalists.

Self-Esteem is Not Job 1
Gentility, politesse, decorum, and high self-esteem are wonderful. You can achieve all of these things within your organization. And then watch it be destroyed by competitors where frank and, if necessary, harsh criticism is encouraged. Technical people, even (and especially) those fresh out of school are always convinced that whatever they've developed, no matter how hare-brained, is perfect. It takes a technical person with good judgement to notice the flaws and it may require repeated and increasingly harsh delivery for the, uh, pinhead to realize his or her mistake.
Example: I once encountered a group of 6 people who called themselves "engineers." To solve what they thought was a new problem, they were going to build their own little database management system with their own query language that was SQL-like without being SQL. I pointed them to some published research by a gang of PhD computer scientists from IBM Almaden, the same lab that developed the RDBMS and SQL to begin with in the 1970s. The research had been done over a five-year period and yet they hadn't become aware of it during several months of planning. I pointed them to the SQL-99 standard wherein this IBM research approach of augmenting a standard RDBMS to solve the problem they were attacking was becoming an ISO standard. They ignored it and spent another few months trying to build their enormously complex architecture. Exasperated, I got a kid fresh out of school to code up some Java stored procedures to run inside Oracle. After a week he had his system working and ready for open-source release, something that the team of 6 "engineers" hadn't been able to accomplish in 6 months of full-time work. Yet they never accepted that they were going about things in the wrong way though eventually they did give up on the project.

An 1994 New Yorker article about Microsoft relates "If he strongly disagrees with what you're saying, [Gates] is in the habit of blurting out, 'That's the stupidest fucking thing I've ever heard!'". Jennifer New, a former Microsoft contractor, writes "Meetings with Bill or one of his top people are often replete with a barrage of expletives and other disdainful comments." (Salon, September 1997) My friends who work or have worked at Microsoft tell similar tales. But how different is this from other elite organizations?

The bottom line: self-esteem is great but beware of creating a cozy home for unproductive people with bad ideas."

Saturday, September 02, 2006

"A genius isn't simply someone with an exceptionally high IQ. Instead, a genius is an extraordinarily intelligent person who breaks new ground with discoveries, inventions or works of art. Usually, a genius' work changes the way people view the world or the field in which the work took place. In other words, a genius must be both intelligent and able to use that intelligence in a productive or impressive way."