« March 15, 2004 | Main | March 17, 2004 »
March 16, 2004
OSBC2004: Clayton Christensen on Disruptive Technologies and Open Source
Clayton Christensen, Harvard Business School Professor and author of the The Innovator's Dilemma and The Innovator's Solution is talking about innovation and open source. He contends that startups fail for knowable reasons. He lists some questions every startup must answer.
- How do we beat the competition?
- Which customers should we target?
- What products will our customer want to buy?
- How should we distribute to and communicate with our customers?
- Which things should our company do and what can our suppliers do?
- How can we avoid commoditization?
- Who should be on our management team?
- What is the best organizational structure?
- How can we know when to change course?
- Whose investment capital will help and whose might hurt?
The "disruptive technologies model" shows that technological improvements in products out paces the customer's ability to utilize or absorb the improvements. These kinds of technological improvements are usually brought about my established companies for their customers. Disruptive technologies start out so bad that they are not good enough for the mainstream market. The trajectory, however, causes the new technology to overtake the established technology. Further, since the existing customers are over-served y established products, this creates an opportunity for a disrupter to provide a "good enough" product at a cheaper price. Established companies almost never bring about disruptive change.
The problem is that management in the established company is faced with the a simple choice: "should we make incremental improvements to our high margin products that our customers tell us they love or should we invest a lot of money in low margin products none of our customers want?" Put that way, management chooses the existing product and the known technological improvements. Most existing giants start out with disruptive technologies, but as they become established, they stop innovating because their customers don't want innovation. He gives numerous examples.
Clayton makes the observation that integrated firms have the advantage when products aren't good enough. The example he uses is the early days of mainframe computers when IBM had a huge advantage because engineers, manufacturing, and software development had to interact everyday, multiple time per day. Eventually systems migrate to modular architectures from integrated architectures. When that threshold is crossed, focused firms have the advantage because they beat competitors with speed, responsiveness and customization. A later example of an integrated architecture losing to a modular architecture is the lose of market by Apple to the IBM PC.
You can't position a firm to be at a specific place in the value chain, because by the time you get there, it will be gone. You have to position the form for where the market is headed. Attractive profits are typically earned in the stages of value-added in which non-standard integration of components occurs. Once the industry moves from an integrated architecture to a modular architecture, the money is made not by the integrator, but by the maker of the subsystems and components.
Clayton introduces the "law of conservation of modularity." The idea is that either the integrated system or the subsystems need to be modular and comfortable in order to optimize performance for the other. He uses Microsoft OS and Linux as examples. Applications in Microsoft are suboptimized in order for the OS to be optimized (this is a result of closed source) whereas in Linux, the OS is suboptimized via an open and modular architecture in order to optimize the application (i.e. being able to change code in the OS to optimize application performance).
In Clayton's world, "modular" implies "suboptimal" or "inefficient." You can't make money on the modular layers in the value-added chain (they're the integators). You make money at the borders to the modular layers. Another way of looking at this is that no one is going to make money on SOAP, you're going to make money on the parts that talk SOAP, the services.
Tim O'Reilly, who Clayton recognizes in the audience and asks to comment, says that Mac OS X is an example of this. FreeBSD is the modular layer in the value-added chain that Apple has put a layer on top of, added value, and is making money there. Clayton says that Motorola and Nokia should supply chip sets and software to Chinese manufacturers who can battle it out in the integrator, commoditized space.
Startups can become established by competing against non-consumption. The example is the transistor radio. It wasn't good enough for anyone who could afford a table-top tube radio. But teenagers embraced it because their option was no radio at all. This allows the startup to avoid the demanding technical hurdle of competing with established products that are better than customers actually need.
06:01 PM | Recommend This | Print This
OSBC2004: When to Not Use Open Source Methods
Martin Fink, CTO of HP's Linux group and author of the book The Business and Economics of Linux and Open Source gave some reasons not to use open source in your business. This was not the bulk of Martin's message, but I thought this idea was good given what I'd just posted about Jim Grey.
- Product is a control point for the company
- Product should go obsolete
- Cost does not justify benefit
- Misdirection and defocusing of resources
- Intellectual property risk cannot be justified
- To compete against open source community
- Just because its cool technology
- Technology direction doesn't match strategic goals
- Time to market is critical
I think that this, seemingly negative, message is a good rejoinder to the statement made by Jim Grey because it stresses that any corporation is going to evaluate open source software on its merits and engage in open source development when its beneficial. While this seems like common sense, it is logic that is lost in the debate. Good, well-managed companies work to increase shareholder value. Consequently, these companies should be engaging in a clear-headed evaluation of the advantages and disadvantages of open source--using it where it makes sense and not using it where it doesn't.
02:11 PM | Recommend This | Print This
Jim Grey Says Open Source Software a Threat to Entire Software Industry
Interesting juxtaposed to what I'm listening to at OBSC:
For-profit software companies will struggle for a business model against free software, said the official, Microsoft Distinguished Engineer Jim Gray. He served on a panel pertaining to software trends, XML, Web services and grids at the Software Development Conference & Expo West 2004 show here on Monday evening. "The thing I'm puzzled by is how there will be a software industry if there's open source," Gray said, disagreeing with a fellow panelist over the impacts of open source.From InfoWorld: Microsoft exec: Open source model endangers software economy: March 16, 2004: By Paul Krill: Platforms
Referenced Tue Mar 16 2004 13:42:02 GMT-0800
Typical of most attacks on open source, this assumes a binary world where everything either has to be open source or closed source. That's not the case. As Larry Lessig just said, property rights are about deciding what you can give away. Statements like the one from Grey seem to cry for some kind of jihad against open source to save the industry. It doesn't have to be that way. No one's forcing Microsoft to open its source. The market may decide it doesn't like that. If so, then so be it.
01:47 PM | Recommend This | Print This
OSBC2004: Larry Lessig on the Creator's Dilemma
The first general session of the afternoon (which in the tradition of modern conferences is called a keynote, along with every other general session) is Prof. Larry Lessig of Stanford. Larry needs no introduction to many people as a champion of the commons and author of The Future of Ideas : The Fate of the Commons in a Connected World and Free Culture: How Big Media Uses Technology and the Law to Lock Down Culture and Control Creativity. Larry is a great speaker and this talk is no exception.
Larry starts off with a brief history of photography and the court decision that was rendered sometime around the same time Eastman introduced the Kodak camera that said "images are free." Without that legal decision, photography would been a much smaller industry. If it had gone the other way, perhaps we would have had a "Daguerreotype Machine Control Act, or DMCA.
Comparing that to today yields a very different result since images, sounds, and other digital goods are not free and in fact are controlled. His point is that this opportunity cost is huge and holding back the market. These unintended consequences are the product of outdated law which is imposing burdens that benefit no one directly.
From 1790-1976, the copyright regime in the US was conditional such that only a tiny proportion of public works were regulated by copyright law. Since 1976 we've had an unconditional regime. No registration, notice, deposit, or renewal is required to get copyright protection. This means that the ability to build and transform upon large parts of our culture requires permission (and hence lawyers) first. This is a burden to the creative process.
This unintended burden is an irrational, unintended burden. There is also a rational piece to this. The rational choice is to be "boring." This is an intentional move by large holders of created works to suppress creative innovation. Larry gives the example of "All in the Family" being offered first to ABC and ending up on CBS as a result of artistic control. In 1976, 75% of television content was produced by independents. Now, 75% of television content is produced by the studios themselves. A big shift in artistic control.
This choice to be boring is a rational step to avoid competition. Larry gives three examples of piracy (including a Peanuts clip, the Grey Album, and a take with Bush and Blair singing "My Love" to each other) that are powerful messages but are at present illegal. This is a democratic potential, but also a creative potential.
Larry gives "science" as an open collaborative model for producing public goods. Other examples include IETF standards, open source software, the human genome project, and the GPS (global positioning system).
There is an IP McCarthyism that is exacerbated by the copyright war. This has created a bipolar world, where you are either for maximal IP protection or you are against intellectual property. (I have to say, that Slashdotters take the opposite position from the RIAA and MPAA in a way that gives credence to this claim.)
Larry now takes on that exact issue: Technologists do nothing to help this debate. He suggests some principals:
- We believe in competition between and within business models.
- Monopolies tend to weaken this position and open systems tend to strengthen this position.
- Our support for open systems is a belief that they are an antidote to the innovator's dilemma and the creator's dilemma.
- Through support of open systems and resistance to the bit-head way of thinking was can regain the Kodak opportunity.
This is not a debate between commerce and something else, or a debate between capitalism and communism, we will not regain the seemingly lost opportunity. This debate is not about doing something other than commerce---its about stopping powerful forces from stifling creativity.
01:09 PM | Recommend This | Print This
OSBC2004: Open Source and Outsourcing
I'm at the Open Source Business Conference in San Francisco. I'm in Steve Korn's talk on Outsourcing and Open Source. Steve works for EDS and has some context on how outsourced IT services vendors are using open source. One interesting note from his introduction: they use Jabber internally and apparently with great success.
Steve talks about the stages he's seen outsourcing go through. In the past, firms outsourced under duress. They totally outsourced their IT services to a firm such as EDS. Next came leverage, function outsourcing. Firms now optimize their internal operations first and then outsource. Steve is starting to see usage-base outsourcing with service level agreements and contracts with declining prices overtime (called forward pricing). Outsourced vendors have to commit to these declining prices contractually.
Outsourcing is being driven by firms that have the majority of their IT spend going to fixed costs, but are being pressured to do more strategic projects. The only way to do that is to reduce costs, get more efficient and drive the fixed costs out to get some headroom to tackle the strategic projects.
Steve did a scan of several hundred EDS clients and discovered some interesting facts. Among medium sized clients, 25% were running on Linux. Government was at the top of the list for usage by sector and then manufacturing. Financial services was last. By region, Asia and Europe were the big leaders. He sees most clients going for a Linux+Windows+Unix solution, but the bias is to stay on proprietary software because they like predictability.
Steve refers to the Microsoft argument that showed that even a few hours of lost productivity for a $70K worker was not worth the risk of open source and responds with the observation that in many countries, workers are making a few hundred dollars per month. This reminds me of being in graduate school and trying to explain to students from other countries why it made sense to buy something instead of doing it ourselves (I managed the Robot Lab). They had a very strong "do it yourself" mentality because they came from cultures where labor was extremely cheap.