Andrew Schmitt, who works at Nyquist Capital has an interesting post on the deployment of optical fiber to the home and China. It shows the importance of scale. Please note that the gchart presented is for global broadband subscribers, not fiber to the home.


Google Q&A

Eric Schmitt, CEO and Larry Page, co-founder were interviewed recently in London. I encourage you to look at this 30 minute clip as it provides a good inside view on how Google is actually run and what their senior management is thinking about these days. Go to the bottom of the page for the interview click. We will be discussing Google at great length next week--this will be the theme for Monday, and on Wednesday we will try to understand how this affects empire of Microsoft.

An interesting story on Intel in China--how they created the "human" infrastructure needed to staff and operate an assembly plant in Chengdu. Please note the accent on human resource development.

Intel Pushes Chip ProductionDeep Into China's Hinterlands
By EVAN RAMSTAD and QIN JUYINGMay 23, 2006; Page B1
CHENGDU, China -- When Intel Corp. began setting up shop on the outskirts of this sprawling city in China's mountainous outback three years ago, Chengdu had a panda sanctuary, some of China's spiciest food and a collection of aging heavy industry.
What it didn't have was a single high-tech factory. Now Intel's plant, which started up this past December, each month spits out hundreds of thousands of chip sets, used to support the microprocessors that are the brains of personal computers.
What Intel learns could help the world's biggest semiconductor company as it pushes its manufacturing further into virgin territory. No chip maker has been as aggressive at expanding overseas. Intel's far-flung network of factories -- in six U.S. states, Costa Rica, Ireland, Israel and three Asian countries -- is key to keeping profit margins high and costs low, which in turn helps lower the cost of PCs that use its chips.
Adapting to Intel's culture of "constructive confrontation," was hard for Star Qin, right, now a team leader in Chengdu.
Intel already operates three plants in the prosperous coastal metropolis of Shanghai. But they are a world away from Chengdu, a city of more than 10 million people in one of China's less developed provinces, Sichuan. Here in western China, factory labor costs average 30% less than in Shanghai, and cities in the region are willing to provide incentives, such as lower taxes, to attract business.
Intel considered several cities in the region, including Xi'an, a well-known tourist center with many more technical universities. When Intel representatives first visited Chengdu in 2001, the land where their plant is now was filled with small farms pocked by hand-dug wells.
But Chengdu left a strong impression on Intel. When the company sponsored a Chengdu computer fair and then-Chief Executive Craig Barrett showed up, 200,000 people mobbed him as if he were a rock star. "We were absolutely fascinated by the passion for technology in Chengdu," said Ian Yang, a marketer at the time who is now Intel's country manager for China.
However, the city's location posed problems. Intel executives worried that local manufacturers weren't the best training ground for its prospective work force. They feared that just-in-time manufacturing processes, which allow the company to turn around orders from customers like Dell Inc. and Hewlett-Packard Co. in hours, would be so new in Chengdu that veterans of other local companies would find it hard to cope.
Many foreign companies in China face a similar challenge, especially as they move farther from the developed coastal areas in search of lower costs and new talent. Some sponsor training programs in universities to help create a pipeline of future employees. Japan's largest advertising agency, Dentsu Inc., built a special vocational school to expose Chinese students to marketing tactics that aren't common in the country. The medical-products unit of Philips Electronics NV is helping to build a biomedical college in the northeastern city of Shenyang.
In Chengdu, Intel decided to train most of its workers from scratch, including hiring fresh college graduates for 70% of its nonhourly positions, compared with 30% in all its other locations. All workers would be sent to train in existing plants in Shanghai, Malaysia and the Philippines. But the training, Intel realized, would have to begin even before graduation.
Intel's corporate trainers visited universities and vocational schools in the region to assess professors and curriculums. To better prepare graduates for work at Intel, the company arranged to add courses on semiconductor physics and factory processes -- and bought the textbooks to teach them -- at two major local schools, Sichuan University and the University of Electronic Science and Technology of China.
"We try to create an ecosystem so we will have enough talent from the local universities," says Yan Chen, director of education outreach for Intel in China.
At UESTC, a former military college converted to private use in the 1980s, Li Bo built a small factory in a classroom, where students sit around a moving assembly line to fit pipes and other plumbing parts together. Cameras and computers record their actions, and the students then look for ways to improve the process.
Recently Prof. Li even began a six-month stint working nights at the Intel plant. "If I know the Intel factory better, I can teach the students better," he says.
The preponderance of under-25-year-olds gives the Intel factory a collegiate air, though Intel's culture of "constructive confrontation" imposes a quick maturity. Star Qin remembers breaking down in tears the first time her work was criticized during training in Shanghai. Her supervisor later told her the criticism wasn't personal. Now the leader of a team of machine operators in Chengdu, she teaches others the same lesson: "At Intel, we point to the problem, not to the person."
When the Chengdu plant has been open for a year, the company's Arizona-based site-selection department will review productivity, turnover, costs and other measurements to see if they meet expectations.
But executives say it ramped up at a faster rate than expected. Meanwhile, two other chip manufacturers, one from China and another from Malaysia, are building test and assembly factories next to Intel's.
Intel is now repeating its recruitment and training process as it finishes work on a second plant in Chengdu, where another 1,000 people will work. And the company has said it will build its next factory in yet another place where no other chip manufacturing exists: Vietnam.

Additional news on China business was the trend setting purchase of Millicom by China Mobile for $5.5 BN. Millicom focuses on LDC cellular telecom and has operations in 16 countries with 10MM subs. Many think Chinese will follow in the footsteps of Japanese and Korean businesses in setting their business model. It is clear that Chinese businesses have their own mind and are creating their own path. Follow it closely. They (as well as the Indian entrepreneurs and business leaders) are quite willing to take risks to shake up their companies and industries--witness Lenovo and IBM. The purchase of Millicom is based on different circumstances, but again shows that they are quite willing to use M&A as a tool to accelerate their development.

I just returned from China yesterday and each time I go I learn something or maybe think about "different" more. Of course everyone knows about the importance of "Guanxi"--relationships. And what is its impact in the way you conduct business and in the texture of business? You hear continuously the stream of complaints about the difficulties that this concept places in the conduct of business--that it can take forever to build Guanxi, that it takes too long to do business there because of it. And of course, China is a big elephant, very difficult to make generalizations.

I view Guanxi as a humanizing element in business--a big positive overall when done correctly. It essentially places substantial brakes on business conduct that can lead to better behaviour. You cannot sell defective products or make uncommittable commitments without ramifications to yourself and more importantly your friend. So people end up speaking more carefully. It places a premium on your personal brand--that you demonstrate competence and trustworthiness. It brings business into what it is--as a part of life rather than an end in itself. Business is essentially transferred person to person rather than Company to Company. And transactions are not a means to an ROI itself, rather ROI is one component of the human relationship. I am sure I am idealizing to a certain extent, but the way in which you do business changes from buying and selling to generating continuously trust and demonstrating competence in your field.

Of course there are excesses, just as there are excesses in the US legal system, which guards well against misrepresentations and unfulfilled contracts. One can see business as a series of trades requiring the fulfillment of commitments and different skill components/products being traded for cash (i.e., arbitrage). The further you move from a strong legal system, the more important Guanxi becomes as a business enforcement mechanism. It is not the opening of the business channel (where most people think the importance lies) but rather the transfer of trust that is the key.

Finally about China and cheating. This happens not just to foreigners but also Chineses domestics as well and it is an issue. One GM of a major business group in China complained to me that they had been cheated a number of times on their contracts with locals. Yet strong Guanxi can help reduce this risk and get people thinking more in terms of win win relationships. To do otherwise is perceived as risky, not just by you but also by your Chinese counterpart. So to do business in China, pay attention to the people, not just to the business.

I was preparing for next week on Apple and came across this interesting article on what lies in store for business over the next ten years. What are the key trends? And it comes directly from some of the innovative business minds today. Though it is not so related to the class itself (though some touch on the impact of the net), I encourage you to read it.

Two interesting articles on telecom, the "whys" of which we will discuss today. The first is that Intel recently announced that it has invested in two WIMAX service providers. And the second was an article that appeared in the UK that deals in fixed line telecom are at a four year high.

Yesterday, we spoke at length about network convergence--the fact that if the services do not get differentiated then we will end up with an industry where the sandboxes come together and the profitability of the sector will decline materially. And new competitors will appear who have brand or a customer base that can provide a good jumping off point to getting into the business. An example is TESCO, the supermarket chain who is rumored to be getting into the Voice over wireless lan (VOIP) business. And why not?

As some of you may know, the latest DoCoMo phones have bar code reading capabilities that are offered through their Felica service. What is the practicality of this? In concept, you can read a product bar code from a package and instantly download product information on the net about a potential purchase. In an open system, you will be able to instantly compare this product to other competing products, enabling you to compare price, quality and other feature fits. I have no idea whether TESCO is considering this capability or not--but technology opens up many avenues to engage the customer. You can imagine putting in place a "closed" system with only information on their products on the shelf made available. So for TESCO in the future, the question may be how to get the handset into the hands of the customer for higher engagement, in addition to the usual "how to build a workable application." This workable application follows the strategy of engagement--not the other way around.

So there can be many reasons to enter the communications business--reasons outside the normal "new business investment parameters. " Another example could be MTV or Virgin Mobile with their mobile virtual network operator strategy (MVNO) as a way to further their interests with their generation set of customers. Businesses are increasingly trying to throw a hidden net around their customers, which technology can enable or disrupt.

OK. The link somehow went afoul.

Here is the letter from the ex-Microsoft employee complements of the NY Times columnist David Pogue:


"Dear David,
"The group you want to think about isn't an insular group of Microsoft employees--it's Microsoft and its OEMs [original equipment manufacturers, the companies that build the computers that Microsoft suggests]. The Microsoft people were probably more aware of the coming failures than the OEMs ever were.
"Microsoft relies on OEMs. And OEMs, in their slim-margin business, rely on a healthy volume of people upgrading or buying new hardware. The tablet PC, the UMPC and the ill- fated Mira (a portable display-only tablet), were joint efforts whereby Microsoft--in the interminable stretches between releases of Windows--attempts to advance the Windows platform AND keep its OEMs engaged, excited and healthy.
"Imagine being an OEM, like Dell, that ships Windows boxes.
'Windows Vista is delayed again for HOW long?', you might think. 'Maybe I should keep up with all this Linux excitement. And by the way, I better figure out how to keep Wall Street and my suppliers happy for another 6-month--12 month--(sigh)--delay.'
"So Microsoft approaches these OEMs with a prototype and says, 'Wouldn't it be great if the market could have something like THIS? In fact, we have a great piece of software for this. All you guys need to do is design one with your own logos and sell it. We'll promote it with you, partners.'
"Now, often, Microsoft really *has* thought the problem through. First, it has years of experience, and failures, to call upon. Second, reference [prototype] designs fill in most of the blanks for a manufacturer. The prototype has probably been usability tested at Microsoft, along with the software, in ways OEMs would and could never spend to do. And reference designs often also include market research that suggests things like a price.
"That's when reality kicks in. First, the manufacturers have to add their own touches, partly for differentiation and partly because they believe that THEY know their customers best. These touches add delays and cost to the effort.
Second, politics often also come into play. When Microsoft sits down with a Dell, an HP or a Samsung, the business of ALL their business come into play--including the all- important Windows licenses--and support for or against something like the UMPC becomes a bargaining chip for both sides. 'You do this, and I'll stand up at your press event and also buy/invest in this standard or that standard,' etc.
"And so, a compromised design or an unnecessary feature gets added here and there, adding time/delay, adding cost...
"Finally, at the other end of all that, there's the issue of follow-through. A few months later than planned, and some factor more expensive than originally planned, the train is rolling. Microsoft puts its weight behind the concept and, correctly I think, keeps pushing for what it believes is truly a winner in version 2, version 3. If you read between the lines, you can tell where Microsoft thinks a product is in its 'maturity' based on which executive announces it, where, and when.
"It's like when a new Robin Williams movie comes out and he appears on the requisite TV circuit put together by the studio. Sometimes he spends his entire interview talking about his movie. Sometimes, even if he's done a great job himself, he spends his time talking about politics or feng shui. We all know what he thinks about the movie, based on the distance he keeps from it.
"I hope this helps you see how Microsoft and OEMs, all acting pretty rationally, sometimes have to take a longer route to get from point A (a good idea) to point B (a good execution).
There are MANY potential pitfalls when one entity controls too much of a design, including cost, insular thinking and inevitable blind spots."

Why do Companies with such smart people do such seemingly "dumb" things? Microsoft introduced a new platform for Windows known as the UMPC. It is supposed to be a handheld PC--lighter than the lightest notebooks and more heft than those little PDA or smart phones. The first product was released by Samsung.......and it looks like it is a clear dud. For a very interesting view of why smart companies can end up looking like they are doing "stupid" things refer to an opinion from an ex-Microsoft employee at the bottom of the article. It is, after all, in the human dimension where successes and failures lie.

We move from semiconductors to telecom next week--ultimately a key theme of the class is about the triangular relationship between networks, devices and content. There is an interesting interview with Terry Semel, CEO of Yahoo on video. I encourage you to take a look. It is a good warm-up to understanding the Google phenomenon which we will be covering later in the course. He talks of the importance of search and advertising; their relationship with Google in the early years and how Yahoo tried to buy them.

From yesterday's session, I hope each of you comes away thinking in more cyclical ways. These cycles are important for taking action (or not) in everyday business decisions. Ultimately business is a step by step move where actions should be plotted out (my opinion) in light of various anticipated event scenarios. And the cycles have much to do with this plotting. From Paul Otellini's presentation you should have noted the micro-architecture cycle--the fact that it took four years for them to refresh their microprocessor core; their technology cycle which has been happening with a regular occurrence every two years; the product cycles (all based on the upcoming new micro-architecture, which are all in the process of being refreshed; their working capital/production cycle which got out of synch with their demand, resulting in excess inventory. And the criciticism of the Company continues as this WSJ article below shows:

Intel Shareholders Voice AngerOver Chip Firm's Sagging Stock
By DON CLARKMay 18, 2006; Page B2
Intel Corp. management received some harsh questions from shareholders about the chip maker's sagging stock price.
One questioner at the company's annual meeting went so far as to suggest that Intel take a page from Hewlett-Packard Co., whose results and stock price have improved dramatically since it brought in Mark Hurd to succeed Carly Fiorina as chief executive.
"I suggest that someone ought to be fired in Intel," the stockholder said during the meeting, held near the company's headquarters in Santa Clara, Calif. "My question is who?"
Craig Barrett, Intel's chairman, conceded that the company's recent financial results and stock price have been "unsatisfactory." But he argued that a more reasonable form of "punishment" for management would come in a reduction in bonuses and stock-based compensation for 2006.
Paul Otellini, who became Intel's CEO a year ago, received a bonus of $2.68 million in 2005, nearly double the $1.36 million he received the prior year, according to the company's proxy statement. His base salary rose to $608,300 from $450,000, and was set for this year at $700,000. He also received stock-based compensation awards valued at $5.9 million, up from $3.3 million in 2004, the filing states.
Reed Hundt, an Intel director who is chairman of the compensation committee, told shareholders that Mr. Otellini's pay fairly reflected the company's earnings growth in 2005. But he stressed that compensation will vary with the company's performance.
Intel's stock, which traded at more than $34 in early 2004 and ended 2005 at nearly $25, has been pushed down lately by the company's slowing growth and market-share losses to rival Advanced Micro Devices Inc. Amid a widespread decline on Wall Street yesterday, Intel shares fell 40 cents to $18.66 in 4 p.m. composite trading on the Nasdaq Stock Market.
Mr. Otellini gave a presentation and discussed the company's plans to make its product line more competitive, as well as an internal study to make the company's internal operations more efficient. But he conceded that many analysts are waiting for signs that Intel's actions are having an effect.
"In the aggregate, they all said 'show me,' " Mr. Otellini said. "Talk is cheap."

Some additional news that we will discuss tomorrow. First, SMIC, China's largest semiconductor company announced that it received $3BN in funding from the City of Wuhan for building a new fab facility.

Second, the Semiconductor Industry grew 7.3% in Q1.

Complements of Kevin....

To Ms. Do, oversupply in semiconductors or LCDs is a common occurrence. Of course, it is not as catastrophic as bananas or tomatoes. In such instances, the inventory value goes to zero mainly because people can eat only so many bananas regardless of the price. In the case of LCDs or semiconductors (especially memory) the vendors just end up putting in more of the stuff into existing systems--going from 512 to 1gb; or 15 inch to 17 inch--all for the same price range.

There were two interesting articles that run with the grain of our class themes. The first is an article by Walt Mossberg of the WSJ. You may remember in the Innovator's Solution that a key is defining the scope of the operation--that companies that develop disruptive "new" things will often need to control as many points of the value chain and system interface as possible whereas modular "cost" based technology companies define their specialty and outsource and rely on technology standards (such as GSM or CDMA, or Wireless Lan protocol 802.11 or Wide area wireless called Wimax or 802.16) or ready-made modules for the rest for the full development of the system. In this article Mossberg compares Apple to Microsoft.


Note that the WSJ is a subscription service. If KDI does not have a subscription, you would do the school good by asking for one. If you cannot get it, let me know and I will send it out.

A second article is on the fact that IPOD phones will be introduced in Japan through Softbank. How do you think the deal got started? Why would Apple choose to start in Japan?

On wednesday I spoke of the importance of understanding the cycles of a business. It is only then that you will be able to feel the rhythm of a company or industry. And there are many such cycles to understand. Let me review a few of them with you:

Industry cycles: These are often demand/supply cycles and we see them most often where supply comes into the market in "big chunks" and where there is little differentiation of product. Examples would include LCD panel production, long distance fiber optic line capacity and semiconductors. The "chunkiness" results in a lot of burps in the industry--though demand continues to grow, the chunks of supply that come onstream outstrip the ability of the end market to absorb. What happens then is what we see in everyday life--15 inch LCD screens give way to 17 inch; 256mb DRAMS give way 512 mbs and prices drop like a rock. Eventually this supply overcapacity gets corrected as companies no longer have the capacity to re-invest--their profits are replaced by red ink--and either they lack the cash or their stock price considerations act as a brake on further investment. Supply then tightens and profits return. The length of the cycle will vary in accordance with the time it takes to add capacity in the industry.

In the semiconductor business this can be as long as 9-12 months to build the fab and an additional 6-9 months to fill the fab with equipment. We have been waiting for four years for an upturn in the backbone telecom industry--those companies that provide backbone lines to corporations and other telecom providers. It has not materialized because these companies laid a lot of dark fiber during the bubble period which can be readily lit and come online--so supply is easily replenished even though demand keeps doubling every year (or less a period) for "bit" demand. Eventually the dark fiber will all be lit and we will be back to the days of shortages--and prices will go up. But when is the magical question and there have been a lot of predictions with a lot of false starts.

Working Capital Cycle: The working capital cycle in a sense is closely related to the above supply demand cycle. As you know, the working capital cycle refers to the conversion of inventory (work in process production and post-production) into receivables and finally into cash. The cycle has become shorter and shorter over time--with management innovations such as just in time inventory management and demand pull inventory systems, which force the vendor to hold inventory. Also (and importantly) better supply chain management tools and software have contributed to the compression of this cycle.

Despite these improvements, the working capital disruptions will still flare up from time to time. Intel, for example, this quarter saw a significant build in inventory which had the effect of depressing margins and revenue. In the disk drive business, recessions in the industry are caused by over-supply which usually shows with too much inventory in the channel. Remember we say that high tech inventory is like sashimi--it declines rapidly in value with time. This only contributes to further pricing pressures in the market. In enterprise software, the time to be wary of the working capital cycle is during new product introductions--the accounts receivable balance is a sure sign of the quality of the product--dissatisfied customers don't pay.

So if you want to know the general health of the industry, watch the supply-demand cycle and the working capital cycle--especially inventory. I will comment later on the other cycles that we see in IT.