Supplemental Project Hunt Report: The New Wind River (Cont.)
Part III: Analyzing The New WIND's Value Proposition
To understand the new WRS, their recent announcements must be appreciated. First, the advanced edition of Tornado/VxWorks maintains twelve years of code integrity, protects existing customers' investment in their application codes, provides a commercial alternative to proprietary HA, involved major partners to ensure fit, and delivers time to market and total cost of ownership benefits.
Second, collaboration in CoEs helps customers develop products in four ways: (a) software optimized for a specific microprocessor architecture simplifies the development process, reduces development costs, and results in cutting edge applications; (b) software available at the time processors are announced permits early prototyping and application development; (c) software that takes advantage of new architectural features allows developers to focus on their own core competencies, and (d) product road maps provide a long-terms growth path and preserve current and future technology investments. These are four impressive competitive advantages; the loss of any of them becomes a switching cost.
Third, the thirty VARs speed product development by pre-integrating complex hardware, removing that burden from the OEM, and by providing reference designs for the product developer. These value-added partners bundle and resell WRS software with their product, usually shielding royalties from the customer and collecting and paying license fees directly to WRS. This benefits the customer by simplifying vendor selection, reducing dramatically time to market, and letting them focus on their core value add. For WRS, it greatly expands business opportunities, moves design win "upstream," produces competitive barriers to entry, and reduces costs of sales.
The 30 VARs include Adobe for postscript printers; Cisco for cable modems; C-Cube for DVD and DVR; Flashpoint for digital cameras; Hewlett-Packard for web-management, Info-Gear (Cisco) for internet appliances, Intel for Intelligent I/O; Liberate for set top boxes; LSI Logic for Digital TV, DVD, digital cameras; Lucent for VOIP; Motorola for digital TV; ST Microelectronics for set top boxes; Telecruz for digital TV, Texas Instruments for DSP processor, and others
Fourth, WindVentures is a new venture capital division of WRS investing $500k to 2MM dollar in startups involved in high growth market segments. WRS is strategically sowing seeds for its future by gaining advanced market knowledge, increasing strategic business opportunities, and adding customers for its software solutions.
These initiatives demonstrate that the new WRS is not just an embedded OS anymore. No other COTS company, much less a homegrown OS can match the new Wind Rivers' competitive advantage: embedded Software Solutions that integrates hardware support, emerging Centers of Excellence and VARs, a fundamentally advanced RASS RTOS, advanced visualization Tools, value-added Tools that yield a targeted complete solution, plug and play stacks, drivers, protocols, and other middleware, and support from 250 service engineers and access to 500 partners providing more specialized products. Over 200 engineering years of intellectual property development create unusually high barriers to entry and an immense switching cost when time-to-market is EVERYTHING in Internet time. Recognizing what is their core and what is WIND's value added context, over 50,000 active developers have chosen the worldwide leader's embedded software solution to increase their productivity.
The New Economic Model. The old WRS was a typical software company who charged: (a) licensing fees for its package of tools, based on the number of designer seats, (b) a fee for services, and (c) a run-time royalty for its RTOS based upon developer's sales of products that included VxWorks.
During the transition to the new Wind River, the company has experienced high upfront costs as it redefines and reorganizes itself as an enabling Internet company. After the transition is completed, it will reap the explosive benefits of low incremental costs.
The current sources of revenue are estimated to be approximately 30-40% from services and 60-70% from products, with tools accounting for 40-50% and royalties accounting for 20-30%. CEO Tom St. Denis's goals for growth are to win more design seats on more projects that produce more design wins that yield more devices that generate more revenues per project while also creating higher ASPs on bundled software licenses (TMS, TIA) and on device royalties designed with value-added products.
For the first time on its analysts day, WIND introduced a model for estimating its hidden economic potential. It used its DSL wins as a case study, whose principles apply to many markets. Between 2000 and 2003, the DSL market is projected to grow 230% annually. WIND has design wins at 47 of the 54 manufactures of DSL products. Run-time royalties are about $1 per unit of sales based upon use of the basic WRS platform. WRS will strive to increase it market share through its reference designs and value-added resellers. WRS will strive to increase its average sales price by introducing value added features, including chip level device drivers, ATM, L2TP, multilink PPP, NAT QOS, VPN, and VOIP. In addition, WRS is developing "Tornado for DSL" (as well as Tornados for cable modems and for set top boxes). Using this strategy of adding value through building in drivers, protocols, firewalls, and the like in Tornado for DSL, WRS expects to increase its royalties by 3-5 times. The estimation model specifies: Growth times Share = Units; Units times ASP = Revenues. In post 7992, Benn reported, using his assumption of number of units and a gradual transition to higher margins, that estimated DSL revenues at $180K this year, rising to $70MM by 2003. In post # 8000, Benn defended the assumptions he used in his spreadsheet analysis of DSL revenues.
Allen Benn's most persuasive argument in favor of "Wind, Going Up, Up, Up!" uses the Lily Pond story to argue that many of WIND's customers are nearing hypergrowth. He uses the phrase "lily pond," or the acronym yalp, yet another lily pond, with the lower-case letters symbolizing classic understated emphasis, as a metaphor to describe a set of conditions. Just as "gorilla" is used as a summary metaphor for Moore's six defining characteristics, Benn is using yalp as a shorthand for "a market segment that enjoy sustainable growth of at least twice the rate of growth rate of the company as a whole, in which the company holds a 50% share, and in which the ASP either is expected to stay constant or even increase."
The largest yalp may be I20 in its InfiniBand reincarnation, then perhaps it is optical switches, followed by DSL, cable modems, and imaging devices. How many more yalps are there? What are they worth? Good financial detective work could model each of these yalps using the WRS model. Allen Benn estimated that run-time royalties for I2O were about $1.40-1.75 per unit. Benn believes the value-added by, say, TMS could increase the ASP enough to create an order of magnitude increase in royalties on RTOS and middleware, but the 3-5 times increase is sure to be conservative because management would prefer to underestimate its future. Thus, the new WRS would have a significant increase in revenues associated with both winning an increased market share and higher ASP's resulting from the value-add of their intellectual property in their embedded solutions software. Although time-lagged, royalties will someday create hypergrowth.
Gorilla Analysis
Shareholder value is driven by value added whenever a company's return on invested capital exceeds its weighted average cost of capital. When excess returns gap upward, and this high spread can be sustained for a long value-duration period, we have an unusually valuable stock. A high rate of excess returns (GAP) and a long competitive advantage period (CAP) result from competitive advantages that can be sustained. How does Wind River's business drive an unusual and sustained level of shareholder value?
WIND's strategic vision is tightly focused on enabling the Internet and its devices: "How Smart Connected Things Think." Their strategy is to amplify four drivers of shareholder value: (a) by providing a broad and deep industry leading product portfolio, (b) by increasing the value delivered in their product portfolio, (c) by leveraging a formidable value chain of collaborative alliances and value-added resellers, and (d) by increasing their penetration of high growth markets. The synergy from these drivers produces four subsequent interrelated competitive advantages: (a) market-leading knowledge products, with high upfront cost and low marginal costs, (b) continually expanding through knowledge investments in the form of R&D and acquisitions that create more value-added intellectual property, (c) that are leveraged by a collaborative value chain that produces a total solution, serves as VARs, and generates indirect network effects that are creating a natural knowledge monopoly, and (d) that, in turn, exponential expand royalty growth for their universal building block that is cascading through multiple market segments, many of which will generate or are generating tornados. Such competitive advantages will achieve lock-in, sustaining Wind Rivers' dominance.
Does WRS Meet the Gorilla Criteria?
Discontinuous Innovation. WIND's discontinuous innovation began when Jerry Fiddler and Dave Wilner recognized the principle that an RTOS and the tools used to develop it could be abstracted to apply to every microprocessor and all product development. By developing a commercial off the shelf RTOS and a set of integrated tools, abstracted as key enabling building blocks, Fiddler and Wilner identified the core function of WRS. Tornado/VxWorks and related professional services offered product developers a discontinuous innovation that provided an alternative to developing diverse homegrown tools and a variety of operating systems, freeing the developer to focus on their own core functions and freeing the industry to standardize on WRS.
Open Proprietary Architecture. WIND's architecture is the coded set of functions incorporated into Tornado 3/VxWorks AE, which are scheduled to be architecturally upgraded to Cumulus next year. It is architecture because it contains proprietary software that is integrated with a microprocessor to form a unique arrangement of parts and functions using an architectural suite of integrated tools with a unique feel, look, and set of functions. It is an integrated architecture that adapts to new markets and requirements through the additions of new or improved elements that are developed by WIND's R&D (HA and memory protection domains), integrated from open source code (TCP/IP), acquired through mergers (EST or AudeSI products), or licensed from other companies (BlueTooth; Java). The openness of the tool architecture is evidenced by the 500 WindLink partners, but the VxWorks kernel is proprietary, remaining uncorrupted, although it is scalable and can be tailored to meet specific needs (IxWorks, OSEKWorks). It is proprietary because the software is copyrighted, and some patents protect the architecture, with seven patents pending for HA and MDP. Also, their architecture embodies 200 years of proprietary engineering intellectual property, and they have unequaled know-how to continue to expand both its connectivity middleware and integration with semiconductors.
High Barriers to Entry. The complexity of 32-bit and 64-bit semiconductors, their variants in FPLG, DSP and more, along with their accelerating rate of upgrades pose a major, if not impossible, hurdle for any new entrants to jump. Instead of beginning with crawling and struggling to learn how to stand, a company would have to be able to dance software-jigs in triple time while jumping from tabletop to tabletop. No existing competitor can port to a full range of chips now. While working around WIND's patents or copyrights, the time and cost of duplicating and integrating a portfolio that could compete with WIND's broad and deep platforms, would require deep pockets and many more than 200 engineering years. Also, could any competitor to find the talent to write the code, much less abstract at a high level in software coding the key embedded processes? The process of forging the collaborative relationships and finding value-added resellers would be a near impossibility to duplicate, given the underlying network effect and Brian Arthur's Law: Of networks there shall be few. The network effects and compatibility advantage of standardizing on the WIND products for the semiconductor companies and VARs are reduced if an alternative embedded software solution (an RTOS in itself, without tools and support is no threat) was adopted. Any existing products that might reside in their best competitors (Microware or QNX) would make overcoming barriers somewhat easier, but no competitor today is close to WIND in either scale or scope. It would be easier for deep pockets to buy WRS, but the purchase would have to be acceptable to powerful collaborative partners; that is, any competitor who hoped to acquire WRS could expect opposition from disadvantaged partners who compete with them.
High Switching Costs. Switching costs are equal to the value lost when forced to choose the next best alternative, usually the offering of the closest competitor. Loss of the best and most complete solution, loss of compatibility across an integrated tools set, drivers, and communication protocols, and loss of professional engineering services create an intolerable loss in time to market and total cost of ownership. Given that WIND's Tornado/VxWork platform is rapidly becoming a standard, the switching cost is the anomie of no standard, meaning no mass market. Also, mini-switching costs aggregate. For example, the loss of several standard solutions, like a needed tool in Tornado, or a specific driver or communication protocol, like TCP/IP for Internet, or of BlueTooth (if that emerges as the short-range RF standard) add up to a significant summed cost. . Often, there is simply no replacement, like for MatrixX or OSEKWorks in automotive controls.
However, the most general appreciation of the magnitude of HSC is gained when you understand the exponential compounding created by the total package. To compute the Total Value of WIND's embedded software solution: (a) multiply the value of the number of semiconductors ported and integrated times the value-added by HA and MPD to the value of VxWorks times the value of the integrated, highly visual, "wizardized" tool set, times the value-added by 500 WindLink partners, times the value added by the expertise of 750 engineers. The sum of these multiplication is the Total Value of the software solution from WRS; (b) Next, from that Total Value, subtract the value of the next best alternative to yield the Corrected Total Value. Alternatively, you could estimate how much it would cost your company to duplicate or acquire WRS. This gives three ways to estimate switching costs: Relative Switching Costs, Duplication Switching Costs, and Acquisition Switching Costs. Next, add the loss of the value chain embodied in the CoEs and VARs as the marketing expense that must be duplicated or added as a switching cost to any of the three estimates of switching costs. Finally, add in the loss created in your company's market capitalization by lost time to market or the loss created if your strategic plan is undermined by your closest competitor who does have access to the WRS embedded solution.
Switching costs are the inverse of competitive advantages, in which value-added become value-subtracted. When competitive advantages create network effects, value-multiplied-exponentially inverts to become value-divided-exponentially for switching costs. Natural knowledge monopolies are built by increasing returns that diminish the returns of competitors, which continually increase the already high switching costs.
"Failure To Switch Costs." Although some "vertically-integrated" companies continue to employ in-house operating systems that were originally developed for their 8- and 16-bit processors, RTOS is a basic building block that is best outsourced. Modularity is a basic principle in the advance of technology. Wind River has mastered the art of software abstraction to create many modular building blocks for embedded systems. The Internet requirements for connectivity ups the odds that no one can remain homegrown for long. Imagine that you are a company with a homegrown proprietary RTOS and a few tools. What does it cost you strategically when you confuse context with core? What are your failure-to-switch costs? Hint: read the BTE and HSC sections above for a few ideas about how to estimate your FTSC costs. Is it worse to reinvent the wheel or to cut off your nose to spite your face? Can you spell "hubris?"
No wonder Sony, who owns the Asperios RTOS, and Sun, who "owns" Java, and Palm with it handheld OS are all moving toward embracing WIND's embedded software solution. What does it cost a company when they fail to use what is becoming the only universal enabling module, the key building block, that helps connected smart devices think? Who has the inside track here? The only winning choice for in-house RTOS companies is to embrace WRS and to refocus on real core competitive advantages.
Strong Value Chain Formation. First, rid yourself of the misconception that the Microsoftian model of abusing your value chain is the best available alternative for managing your customer-partners. Instead, notice the immense value generated by WRS's collaborative partnerships through its CoEs and its collaborative reference designs with its 30 VARs. This is how intellectual property is expanded exponentially. These partners ARE WIND's CUSTOMERS. As partners in a complementor relationship, each gives to both. Their joint contribution adds necessary value, each provide a building block that together co-create a total product or solution. Then, WIND's partner simultaneously becomes its VAR and Customer. Since when do you compete with and seek to control your customers? Competition yields zero-sum outcomes. Whereas, collaboration-as-virtual-integration can generate orders-of-magnitude improvement in win-win outcomes. Think of the necessity of pre-integrating chips and software because software is the "last mile," the "sticking point", the "final" problem to be solved in technology's advance.
What does it mean when WRS tells its analysts that it has "moved upstream" in its value chain? This is a master strategic move. First, it means that your partners become your primary customers. Think of how you want to make these partners into satisfied and loyal customers: add value, add value, and add value! (It is O.K. to enjoy the lowered costs of selling that this creates, but no gloating is permitted. Save the gloating for late at night when you realize the demand-side economies of scale that this collaboration creates.) Second, when you move upstream and sell to your partners then these downstream VARs sell to OEMs further downstream who sell to customers. Now imagine that you are upriver and your downstream partner has an OEM customer further downstream OEM customer, who has these downstream mass-market customers. Wow! Then, remember that you are establishing many such cascading complementor relationships. Imagine a river is branching in a fractal pattern. Double Wow! Even more amazing it is an inverse fractal pattern from that of gravity-based rivers because its metaphorical branching is branching as it goes uphill, rather than flowing downhill from the mountain top rivulets into creeks into streams into a flowing river as it meets the sea. Let's imagine that instead of a sea at the end, we have WRS as the origin of this fractally-diminishing river. Then you can see the fractal pattern flowing from this huge WIND RIVER that divides into streams that divide into creeks that feed tiny branches that flow into LILY PONDS. Can you imagine that? Wind River is branching and re-branching and flowing into many lily ponds.
Tornado Market Extant or Foreseeable. If you can take your eyes off the compounding green biomass in the lily ponds long enough to turn your attention from green to whether a tornado is extant or foreseeable, you may discover your own green answer. Green. If you can't see the green lily pads doubling, wait until the last green week of summer. Or, recall that exponential compounding is always invisible at first, and place your buy. Green.
As a natural knowledge monopoly-in-process, the dominance of Wind River is inevitable. Does WRS have a sustainable competitive advantage? What were those BTEs and HSCs and FTSCs? Does WRS have a downstream mass market generating run-time royalties? Does WRS have demand-side scaling? Does WIND own the only universal building block in the Internet and its devices? You may differ in your judgments about whether lock-in is extant or foreseeable. Only those without imagination will not foresee an extant lock-in. Can you connect the dots? So far, the analysts don't see the big picture. Talk about a large exploitable gap in expectations for a knowledgeable investor to take advantage of: Wow, wow, and wow!
Advertisement: Stay tuned to this thread for a tentatively titled sequel: "Gone with the WIND" or "Are Network Effects Truly Monstrous?" or "Answering Merlin's Questions."
Disclaimers: I am long WRS. Among other things, the "wow's" are just my opinion. I am intellectually responsible for any errors in fact and for all of my interpretations. I am intellectually indebted to Allen Benn's posts on his Wind, Up, Up, Up! Thread. I am trying to give back some of what I have learned from all of you. This is no substitute for your own due diligence. I am indifferent to whether you are long, short, or indifferent to WRS. I am not indifferent to your reactions to this post and would appreciate any comments you care to make publicly or privately.
I hope this helps.
Don Mosher |
