Post a Comment
News and New Products
Special Report: New Business Models
By Ed Sperling -- Electronic News, 8/10/2005
Business as usual is hardly a term that applies to the technology industry, but even the innovative business models that have created Silicon Valley are undergoing radical changes.
Despite the training that most of them received in business school, CEOs of major electronics companies now equate their roles more closely with managing portfolios or in-house conglomerates than leading hierarchical corporations. Those portfolios can include anything from individual business units to intellectual property, and often they extend outside the boundaries of companies to include partners in a global ecosystem.
Portfolio management has another kind of benefit to a company’s bottom line, as well. It’s easier to determine which groups are doing well and which ones aren’t. When one segment of the business doesn’t deliver on growth expectations, it can be thoroughly evaluated to determine everything from strategic and economic value to market competitiveness -- and whether it’s worth keeping, selling, or writing off.
That kind of thinking makes it tough for rank-and-file employees to think about a long-term future with a company. In their eyes, it’s often viewed as the culprit behind offshore manufacturing and design. But from the top looking down, it makes it far easier to assess where to improve efficiency and competitiveness within a company, whether that includes in-house or offshore manufacturing, or a combination of both. It also makes it easier to determine where money should be spent, whether research is done collaboratively with a group of partners or independently. And the tighter the supply chain integration, the easier it is to figure out whether to buy, build, or weigh in at any phase of a project.
In many respects, this is a marriage of internal and external capabilities, and it can have a huge effect on bottom-line results and time to market.
“I really manage an IP portfolio,” said Wim Roelandts, president and CEO of Xilinx. For his company, that includes three separate layers of IP. One is the IP that Xilinx puts on its chips, which includes things like programmable logic, gigabit transceivers, memory and DSP blocks and clock managers. “Each one of these blocks has its own design group, so each one evolves that core. If you want to develop an FPGA, we provide the connection points. This is silicon IP.”
The company also develops horizontal IP, which includes DSP processing and backplane interconnect IP. “This is IP that is soft logic and which sits inside the chip and allows you to take advantage of these hard cores. The reason we separate them is this type of IP is used in multiple markets. The third level is the vertical market IP. If you go into a base station, it uses certain DSP cores, a backplane and an embedded processor. This is almost like middleware. You can take these three pieces of middleware, two pieces of the backplane middleware, and put it together. That’s not the total base station, but 50 percent of the base station is already there.”
In addition, Xilinx offers its own IP for sale, or lets customers or third parties provide it. “It’s really kind of an assembly. You have the strip IP, horizontal IP and then industry IP.”
In the case of large companies, much of that IP is developed across various business units within the company. IBM, Texas Instruments, Intel and Freescale, among others, draw from across business units to create individual solutions for customers seeking individualized solutions. And while all of them do work with partners when the opportunity is in their best interest, it’s always in the best interest of smaller companies to match external partnerships with their core offerings. Large companies, in contrast, can simply assess what they have available and make use of all the pieces.
TI developed its OMAP platform for consumer handsets at the request of its top Japanese customers. “They said, ‘Here’s the video and multimedia capability we want that device to have,’” said Richard Templeton, TI’s president and CEO. “In some ways, our guys thought the requests were crazy. We said, ‘Okay, they may be crazy but they’re the customer.’ We reached across the product areas inside. The wireless team needed an imaging engine for the video modes. We have a digital imaging unit within TI, and so it comes directly across the borders. You don’t have to go outside and acquire. We announced that wireless LAN at CTIA. That was driven out of our broadband business because wireless LAN is the first use of this technology. It’s to deliver broadband access within a home. Yet the rate that wireless LAN looks like it’s going to be added to the handset is accelerating rapidly. We introduced a dedicated wireless LAN device that will go inside a handset, tuned for power, tuned for space, tuned for chip size and cost. We didn’t have to go outside TI to be able to gain access to that really complex intellectual property.”
TI did the same with its high-speed SERDES (serializer/deserializer). “We can do those in our ASIC business, we can do standalone chips on those, and we have road maps for those serial I/Os that are going to go onto our high-performance DSPs. But that stuff is all internal.”
At Applied Materials, President and CEO Michael Splinter views his role as more of an in-house conglomerate manager -- a somewhat different model than companies like TI because Applied serves a much narrower customer base. “We’re really in one market, and we’re very focused on one set of customers and providing solutions to them. We want all of our business divisions to play together on the internal development of machines and in the application of the machines. This is where our Maydan Technology Center comes into play. We try to make sure our machines work together very well so they etch a low-k film better than the competition can etch it or our metrology can measure the defect produced in film better than anyone else can measure. We work very hard on trying to get that interrelationship on the application, as well. They don’t get separated out even though they might be in the same spot, which is what portfolio management would be.”
Nuzzling Up to Customers
Both models serve the same purpose. Templeton says the real key is what many industry executives have been talking about for years, namely getting closer to their customers and trying to understand their needs.
“It’s still a matter of find good customers, take care of good customers, and let good customers drive you to do competitive products,” he said. “Don’t get confused with internal strategies and ones that are truly connected to customers. That sounds obvious, but when you look at the number of semiconductor companies trying to define grand strategy, it leaves out a pretty important element. Do you have a customer that you’re taking really good care of? That will determine whether you have competitive IP. You then let that drive critical IP.”
Sharing across business units is key to that strategy for large companies that have those capabilities. “Clearly, you need a little more of a re-use view, and you’ve got to have IP be able to move across business units,” Templeton noted. “You can’t keep stuff captive within a business unit. But in the end, we’re still believers in customers and business units.”
He’s not alone in that approach. The mantra among every technology company, consulting firm and business school these days is building partnerships with key customers. In many cases, that includes customized offerings, no matter how complicated the product. And just as often, it means involving customers in the development process -- something that is far easier with a portfolio or conglomerate model.
“The power of a portfolio is being able to structure a solution that draws upon the entire breadth of the company,” said Sumit Sadana, senior VP of strategy and business development at Freescale Semiconductor. “A lot of customers are trying to cut time to market. At the same time technology is more complex. They’re looking for more of a solution from us.”
He noted that selling of this strategy also is different. “Historically, we had sales forces from three businesses. We’re transitioning to a single integrated sales model rather than a single-product driven sales approach. It’s an account-driven sales approach rather than an industry-driven one. There are not that many companies that can bring this solutions-oriented approach to the market. In the case of smaller companies, they will be driven by the partnership model.”
He said the advantage larger companies have is being able to use their own products rather than what’s available for mass consumption on the open market. “The benefit we have at $5.7 billion in revenue is that we can do embedded flash, not flash. This is all IP-driven value-added memory, not commodity memory.”
Rush to Success
Key to this entire strategy is speed and communication both within a company and across and ecosystem of trusted partners. It’s all about the time it takes to get complex products to market first, and that requires an extraordinary ability to pull together all the necessary pieces quickly enough to capitalize on new market windows.
How effectively this model translates into long-term growth remains to be seen, but the ability to respond quickly appears to be a priority everywhere. Perhaps the most extreme example is occurring among startups in China and Taiwan, where original design manufacturers are combining design and development with manufacturing. When a new opportunity crops up, many of these companies change skins to become whatever the customer needs, whether it’s a manufacturer, design house, fulfillment house or a manufacturer’s representative.
“When we send out a piece of equipment, it’s different for almost every customer,” said Applied Materials’ Splinter. “The base machine is the same, so we’ll ship an Endura to almost every player in the semiconductor industry. But most of the major players have specific modifications that they’ve made or that they request to be able to do specific process that they need done in a specific way. Our industry has been one of customization for a long time. The difficulty in that is it can drive up costs, so we’ve been working hard to offer customization where it really matters to the customers and drive standardization across all of our platforms where there may still be customization. But it will come off a standard base.”
Splinter noted that with gas panels, customers want different mass flow controllers and gas panel configurations. He said the solution is to offer a modular panel that provides very high volume at minimal cost, while still allowing customers to configure it the way they want. “You’ll see us do more of these standards-based things over time.”
In the case of companies like Intel, IBM and TI, which often set the standards for the electronics industry, more companies will likely be forced to play in their camp as the cost of development and production goes up. The only question for them is where they will play in the ecosystem of the giants and how they can improve their leverage.
This article is part one of a three-part Electronic News Special Report. For more of this week’s coverage, click here. Stayed tuned to Electronic News for continuing coverage of Special Report: New Business Models in coming weeks.
