Tuesday, February 20, 2007

What Do Zero Switching Costs Mean?

I recently read an interesting article about the impact of switching costs on IT and consumer buyer behavior. Switching costs are the economic and non-economic costs of switching from one thing to its alternative. For example, the cancellation fee charged by mobile phone carriers are an economic cost to switching carriers before your contract has expired. This is clearly an effort by carriers to reduce their subscriber "churn". Alternatively, carriers will want to protect their investment in a subscriber's subsidized handset.

There are also a variety of non-economic switching costs which influence buyer behavior. For example, re-learning how to use a device can be a switching barrier. This is often cited by corporate IT departments in their delay to transition to new operating systems such as Windows Vista. Another non-economic switching cost is an investment in related infrastructure. For example, software written for Sun's Solaris operating system will not run under Microsoft Windows or IBM's AIX operating system, stranding those hardware assets in a transition. Similarly, a consumer switching from standard definition DVD's to Blueray high definition DVD's will not be able to play back a completing HD-DVD format (except for some dual format schemes which been introduced).

In general, switching costs favor the incumbent player and are a barrier to entry for new players into the segment. For example, anyone with a large library of music purchased from iTunes cannot play back this music on non-Apple MP3 players because Apple's AAC digital rights management software is only found on iPods. Other reasons why switching costs favor the incumbent include behavioral inertia, stranded assets, risk aversion, and lack of complementary products. So, any new product entering an established segment must demonstrate a clear and obvious user benefit in order to overcome these switching costs.

Many incumbent players use switching costs as a way of maintaining their market share (think of cable companies, etc). What happens in a world of zero switching costs? This trend is starting to happen in Internet software both in the enterprise space where "software as a service (SaaS)" is an alternative to "shrink-wrapped" licensed software. A more immediate example is in on-line consumer services such as YouTube, MySpace, etc where subscribers have a choice of of competing on-line services. How do such services both retain their existing customers and gain new ones?

This is a new paradigm for marketing professionals to address.

Wednesday, February 14, 2007

What's Next for Moore's Law and its Impact to CE Devices

Recent announcements from Intel and IBM about new breakthroughs in silicon fabrication technologies have poised the question, how does this affect the CE space?

First, a brief summary on Moore's Law. Gordon Moore, a co-founder of Intel, observed the trend that transistor density was doubling every 24 months. He postulated that this trend would continue in the future. This is not a "natural" law like the First Law of Thermodynamics in that it is based on various technology improvements rather than physical phenomena. Nevertheless, the "law" has continued to hold true due to many breakthroughs in semiconductor design and manufacturing.

As a result having twice as many transistors to utilize every 18 or 24 months, designers can either add more features to their die or keep the transistor count constant and reduce the die size by half, and thereby reduce the manufacturing cost, or do a combination of both. The latter accounts for the unparalleled price reductions seen in semiconductors. The trend for the former was to add more complex features to the microarchicture of the processor such as out-of-order execution found in the Pentium Pro processor, hardware multithreading and larger on-die caches. Now, the trend has been to add multiple processor cores to each die to increase the on-chip parallelism at a thread level rather than at the instruction level.

Here are my thoughts and personal opinions on the impact to consumer electronics:

1. Extra transistor budget will be used for more integrated System-on-a-Chip (SoC) designs and eventually multicore processors. More highly integrated devices will reduce size, cost and power consumption of the product. Many consumer electronics devices today are not constrained by processing power but the trend toward processing high definition video content requires about 8x the bandwidth and storage as standard 480i video.

2. Some breakthroughs announced on reducing transistor leakage current and supply voltage will translate into lower power and longer battery life devices. This is a key area in improving the utility of these devices for consumers.

3. Improved functionality and utility of the devices enabled by semiconductor improvements. For example, large solid state disk drives using Flash memory instead of magnetic disk drives for portable media players such as Apple's Video iPod. Or ubiquitous Internet access embedded into the device using WiFi, 3G, or WiMAX radio access.

Just as semiconductor technologies have enabled today's mobile phones, digital cameras, and personal computers, new categories of consumer devices will be enabled by these developments in the future. It's up to us to figure out how to make them.

Monday, February 12, 2007

The New Soul of a New Machine

While I was on vacation a few weeks ago in Panorama, British Columbia, I had a chance to re-read a favorite book, Tracy Kidder's Pulitzer Prize winning The Soul of a New Machine. Originally published in 1981, it describes the engineers who developed Data General's Eclipse MV/8000, a 32-bit minicomputer. I had originally read this book in the summer of 1983. I had completed the sophomore year in the engineering program at Harvey Mudd College and was working for the summer at TRW's Space Systems Group when a co-worker recommended the book.

I was completed enthralled by the Kidder's description engineer's passion in creating the "Eagle" computer and resolved to change to redirect my studies to computer engineering. I've been working on marketing, developing, and selling computers ever since.

What I was wondering was, what is the equivalent of the Soul of a New Machine today? The superminicompter segment that Kidder was chronicling in the book is no longer around today, replaced with lower cost, microprocessor based servers on the low and midrange segments and mainframe computers (still around!) at the high end. While there are still a few new microprocessor designs around such as Intel's Itanium processor, Sun's Niagra design and IBM's Power line of processors, for the most part, developing new computers is kind of like assembling a complex Lego set with parts from various vendors, albeit much more involved.

Is today's Soul of a New Machine a hardware/software solution product like the Apple iPod, Nintendo Wii or Playstation 3? Or is it some types of Internet-based Web 2.0 applications?

What do you think is the today's Soul of a New Machine? Post your ideas here.

Saturday, February 10, 2007

Ultra Mobile PC's and Ultra Mobile Devices at CES

I've finally gotten around to writing about an area which I'm working working on, developing ultra mobile products. These are small portable devices which add ubiquitious Internet access and other functionality to consumer electronics products.

Here are some links to coverage of Ultra Mobile devices from CES:

Engadget has a good summary of Ultra Mobile PC's at CES.
Here are three ultra mobile devices which I am currently leading development teams for:

This device is a next generation Ultra Mobile PC which was first unveiled in Intel CEO Paul Otellini's keynote speech at the Fall Intel Developer Forum, San Francisco during September 2006. This design incorporates a swiveling display to reveal a thumb keyboard below it. It is designed to provide consumers a portable way to take the Internet along with them wherever they may go.




A second design is designed to be used by middle school students ages 11-14 as lightweight, rugged portable learning PC to be used in the classroom, home and in-between. It incorpates a number of features to improve the surviability of the device while being carried around. One design innovation is the padded fabric cover which serves as a protective cover, a hinge for the display and keyboard pieces and a stand. This device should not to be confused with Intel's Classmate PC project which is meant for developing markets.

This third design is a concept of an ultra mobile device designed for car navigation and entertainment but which can be removed from its car dock and used a portable Internet terminal.
I'll continue to add more comments about these and other devices as they get closer to introduction.