<< Finally, one small rant: I appreciated so much of Allen Benn's commentary regarding WIND and its potential opportunities. I find it very perplexing that when the whole tech sector imploded that he disappeared. It gets under my skin that when WIND hit this major turbulence he became silent.>>
Allen Benn has disappeared from SI for nearly 9 months now, so it is unlikely that his absence is due solely to disenchantmenw with WIND. Other personal factors are likely to be at work.
Here is his last post:
<< To:David Swart who wrote (9816)
From: Allen Benn
Monday, Aug 13, 2001 2:29 PM
View Replies (2) | Respond to of 10060
I personally have come to the conclusion that the software tools business is a difficult business to prosper in even if you are the market leader. I think it is much better to make money selling software applications. This particular business is an Alice in Wonderland situation where you are forever running (developing new technology) to stay in the same place. The future prosperity that is always alluded to will never come.
Anyone care to disagree with that and tell me why history will not repeat itself?
Embedded systems is the sum total of a vast number of market segments. Most of the ones that have gotten off the ground are still very much in early stages of development. According to Christensen (of ?Disruptive Technology? fame) products tend to mature in a consistent pattern: When the product is first introduced, integration occurs at the level of the product as a whole. The manufacture strives to meet customer demands by optimizing each component in concert with all the other components comprising the product. In this phase, suppliers to the manufacturer have little pricing power as they are forced to deliver non-standardized components with little influence on the overall product design. The OEM controls the value-add, and therefore the lion share of profits. As customer needs become fully met and the product matures, continued integration at the product level suffers diminishing returns, outweighed by the benefits of modularization and specialization made possible with standardized interfaces. The functions of the OEM gradually shift from being chief integrator to assembler of parts. With the shift in function comes the shift in relative value-add. By the final stage, the OEM is stripped of any economic advantage, with most of the profits going to dominate suppliers.
On this 20th anniversary of the IBM PC, almost everyone ridicules IBM for outsourcing the OS and the microprocessor, the two components that now claim the lion share of profits in PCs. However, at the time those components were outsourced, IBM was in control of all the integration and most of the profits, simply outsourcing the non-core, low-value-add parts to suppliers ? exactly what business schools taught executives to do then, and what they still teach executives to do today (See PostScript for interesting exception).
When WIND first got an OS contract from GM, automobiles took five years from design to production. After nearly 100 years, automobile companies still controlled most of the design and integration of cars, including electronic components that were just beginning to pick up steam. Today, just five years later, design to production can occur as fast as 18 months, with electronic components claiming about 15% of the total costs. As car companies are forced to incorporate standardized electronic control networks and acquire standardized electronic parts (like fuel injectors, automatic braking systems, drive by wire, radar, etc.) they are rapidly becoming assemblers, not integrators. The OSEK standard for networking electronic controls in an automobile is an outgrowth of the industry, not of any single manufacturer. As assemblers, they not only are destined gradually to lose control of how cars integrate parts, well-positioned suppliers can begin to take a greater share of profits. In 1996, GM and all the other automobile companies were steadfast against paying royalties on Intellectual Property (specifically royalty on RTOSs). That steadfastness cannot long stand against economic reality. As with all product development cycles ? even a hundred-year-old product like the automobile -- modularization implies a shift in economic power favoring the specialized provider of standardized parts.
The cell phone is maturing rapidly, already taking the form of the assembly of standardized modules. Qualcomm surely saw this when they exited the handset business. Motorola most assuredly sees this as they restructure their cell phone business enabling them to sell components to competitors. Ericsson wants out of the handset business and Nokia must fear the imminent arrival of Chinese handsets ? with Motorola parts.
WIND has played seriously in the communications sector for a number of years, but prior to TMS, its role was as a commodity supplier to OEMs burdened with all the problems of product integration. I doubt WIND?s engineers had the foggiest notion of what goes on inside a carrier-level circuit switch, even though VxWorks was commonly used in the line-card component of such switches. Nowadays WIND?s management emphasizes the ?integration? not only of all the hardware and software but of the functionality underlying managed Ethernet switches, providing OEM?s risk-free and rapid time-to-market. They also receive three to five times higher ASPs for providing integrated solutions and reference designs. What is really happening is that WIND is becoming the specialized provider of key, standardized components, the first step toward shifting control, and ultimately most of the profits, from the OEM to the supplier.
By any measure, network equipment is ripe for a shift to standardized components, with OEMs reduced to, or replaced by, assemblers. An ominous sign of this was Dell?s recent announcement that they are entering the network switch business. Dell?s recent emphasis on storage products is just as ominous for the high-priced, highly integrated storage system makers, like EMC. Incidentally, according to Intel?s webpage, Dell uses iRAID building blocks for standardized components. Intel?s Intelligent Network Processor family, with all the other building blocks, are positioned specifically at network equipment and its lack of specialization. Intel clearly wants to modularize the sector and pocket most of the profits. As Intel?s unheralded software partner in this exercise, WIND wants to share in the re-allocation of profits.
WIND?s business model is still dominated by the role of supplier to OEMs controlling product integration. It is beginning to change with vertical solutions and integrated reference platforms, but obviously not enough yet as a percentage of overall revenues to reshape economics of the business. That will change over time. I just hope it can happen in less time than it has taken the automobile industry to go to the next level of modularization after Henry Ford?s famous introduction of the assembly line into automobile production.
Modularization usually occurs over a few engineers? dead bodies, but it is an inevitable consequence of competition. It follows, therefore, that the shift occurs most rapidly during times of financial stress, like in the current global economic recession. It happened in the US military only when the cold war crushed the military budget forcing a shift from Mil-specs to COTS components.
Paradoxically, if economic stress speeds up the modularization of products, why haven?t these pluses balanced more of the minuses WIND has suffered in this slowdown? You might recall I argued previously that there are pluses and minuses in this slowdown for WIND, specifically mentioning outsourcing as a plus.
I believe the answer is the following: The minuses occur almost immediately, the minute companies begin to sense the slowdown. The pluses are equally real, but their effects are lagged. When large companies like NT, Alcatel, Philips, Cisco, Lucent, etc. fall off a revenue cliff, their first action is NOT to start outsourcing everything in sight. Their first action is to announce a restructuring with layoffs, and then engage in a corporate-wide study to figure out what core businesses they should retain, and what non-core businesses they need to outsource or even exit. In the meantime, product development and associated procurement are delayed or, at best, limp along. This period probably lasts up to a year or even longer.
I know of one very large company that reportedly just completed the obligatory study, only to discover that 40% of development engineering hours are spent in ?integration? ? most of which is an outgrowth of the increasing complexity of software. Now we know from Christensen?s theory that this is acceptable only if the products require high-level integration. Otherwise, savings of up to 40% of engineering costs are too significant to leave on the table, unnecessarily burdening the company?s cost structure. In this global economy, management has no choice but to take software modularization to the next level in order to remain competitive. Essentially this means outsourcing platform integration and buying into vertical solutions. It is no coincidence that this company came to this sweeping generalization while in the midst of a major restructuring caused by the global slowdown.
Now you also know what?s behind WIND?s increasing emphasis on ?integration? in their discussions with analysts and customers alike. Integration, as distinct from assembly, is the key determinant in the allocation of value-add and profits. This is not to say that management necessarily is even aware of Christensen?s latest musings. The dominant forces of change occur because of Darwinian-like evolution occurring naturally in corporate structures, not because the Harvard Business School invents them. (See the PostScript)
The peaking of the PC paradigm, WIND?s growing dominance in embedded systems, WIND?s impressive strategic relationships with semiconductor companies on the one hand and major OEMs on the other, all played out over the last five years exactly as predicted. That so many outcomes happened virtually on cue is remarkable, given that WIND was just one $20 million company among many and ?post-PC? had not even been coined. However, until WIND shows concrete signs of winning the integration battle by earning a greater allocation of product profits, the final measures we seek, recognition by the Street of expected robust revenue and earnings growth, will remain just out of reach. Everything seems to be in place for this transformation ? WIND?s last great hurdle -- but the proof is in the numbers, not theory.
Allen
PS - I noticed that another Harvard Business School professor, Porter (of "Competitive Strategy" fame) sees oursourcing through different lens when he commented about the slowdown for the current Business Week magazine:
?The short-term cost savings of outsourcing were very apparent, very attractive, and very seductive to companies who were desperately trying to improve their earnings per share quarter to quarter. But when you outsource something, you tend to make it more generic. You tend to lose control over it. You tend to pass a lot of the technology, particularly on the manufacturing or service delivery side, to your suppliers. That creates strategic vulnerabilities and also tends to commoditize your product. You?re sourcing from people who also supply your competitors.?
Actually, Porter?s view on the consequences of outsourcing (modularization) is entirely consistent with Christensen?s. If they differ at all, it is that Christensen points out that modularization (outsourcing) is an inevitable consequence of product maturation, while Porter seems intent on warning companies to outsource (modularize) less than previously recommended, in an (desperate?) attempt to avoid the consequences.
>> |
