How does a product go from bright idea in a single mind to wildly successful global product?
This short essay is not intended to be a complete infonomic analysis. Instead, it is designed to show a qualitative case study where business flow channels may be amenable to quantitive analysis based on the Constructal Law. In addition, the early adopter strategy analysis could be applicable to a wide range of business cases. A more detailed exposition of the fundamentals behind this case will be published in due course.
The Constructal Law and Infonomics potentially provides an analytical engine for understanding the superimposed flows of information and objects as they spread across the landscape from first invention to ubiquity.
As a product developer for many years and innovation promoter as co-founder and CEO of the Silicon Valley World Internet Center, a “Think tank, collaboration laboratory and showcase for Internet invention,” I had the opportunity to observe and study the innovation process, both historically and in real time across hundreds of companies, from startups to mega corporations.
In this short study, we will introduce the Constructal Law as a conceptual method for characterizing processes that were previously only understood through disconnected intuitive interpretations.
The effect occurred on a large scale in the early days of the computer industry, as documented by AnaLee Saxenian in her 1996 book Regional Advantage. Briefly, the East Coast computer companies, DEC, IBM, Data General, Unisys, etc. followed the traditional top-down business model of long-established large companies. Their vertically integrated manufacturing processes fed into a long and fast distribution channels aimed at a limited number of major corporate buyers. What IBM built, companies would buy, without question. “No one ever got fired for going with IBM” was a well known mantra.
Under the autocratic rule of Ken Olsen, Digital Equipment Corp. designed and built every part of their computers, from power strips to software to processors. An employee who left the company was likely to be blacklisted as a traitor and might find it difficult to be hired by other companies. Evolution was only allowed to come from within the walls of the organization. An entrepreneur striking out on his own would find that he could not buy even the simplest of parts to create a competitive product. Mr. Olsen famously threw young Steve Jobs out of his office for suggesting that anyone would want to actually have a computer in their home, not to mention that Jobs then told Olsen, "We will bury you."
On the West Coast, the invention of the integrated circuit by Robert Noyce and others, was the primary driver of innovation. By nature, the IC is not a finished product, but can add value to many products. Rank amateurs, inspired by the easy availability of surplus parts began building their own computers in the famous Home Brew Computer Club, which included Steve Jobs and Steve Wosniack.
Engineers working for competing companies would meet after work and freely discuss their ideas and progress. Employees were encouraged to switch companies and return later. Failure was celebrated as a badge of honor and actually built credibility with investors.
Intel, the most successful maker of ICs and microprocessors, does not itself make a finished product. Instead it sells its chips to any and every company that wants to market a computer product, even raw startups with only a glimmer of viability. Horizontal integration via multitudes of short and slow channels with a few large and fast ones.
One of the best examples of the early adopter strategy was HP. Starting out of a residential garage, they made specialty instruments for engineers. Their market was individual engineers who spread the word through their personal networks. Short and slow, but ultimately vascularizing beyond the engineering community and leading to a huge success.
The characteristic of these short and slow early adopter channels is that they have a “fanout” that depends on the individuals involved, which starts slowly but grows rapidly over time. It can be agonizing for an investor to see growth from 10 to 100 to 1000 in the first year, knowing that year two will bring only 10,000 and the next year, only 100,000. Of course once you get on the steep part of the S curve, it seems like the sky is the limit, but investors are often impatient and want to see sales of a million in year two.
Internet in Oregon
My proposed study example is from a project I participated in around 2002. Based on my experience developing the markets for early cable modems, a venture investor asked me to help out with a company he was funding to provide wireless Internet services. The company, located in San Francisco, we will call Vnet. They were pioneering the use of cellular telephone technology for Internet provision. Their “base stations” were modified cellular transmitters and the Internet “receivers” located in the home looked quite a bit like common wifi routers, but with more powerful radios.
By serendipity, in the 1970s, my wife had worked for an entrepreneur (we’ll call him Cliff) who made his fortune acquiring cell phone “spectrum” licenses when the government was allocating them by lottery. Each chunk of spectrum would allow an operator to install cell towers and sell service within some geographic area.
It turns out that Cliff had acquired a number of licenses in and around his 2,000 acre ranch outside of Klamath Falls, Oregon and he also owned the local “signal hill” which is the optimum place to put radio transmitters. Almost every community has some version of a signal hill, which you can spot by the antenna farm at the top. Cliff had a perfect setup for implementing the early adopter strategy.
Meanwhile, the major cell carriers were battling it out like godzillas within the large cities, with every scrap of available spectrum being bought, sold, traded and fought over for megabucks in expected revenues.
Cliff’s marketing area consisted of a diffuse mix of retirees and farmers living in widely spaced ranches and micro-communities of around 100 or less. Spectrum was available, cheap and Cliff owned it. The strategy we adopted was a copy of the one used successfully by Craig McCaw to become a cell phone billionaire - start small in the communities the majors don’t care about, spread across the landscape and when the godzillas come to fight over your holdings, sell out for plenty.
Vnet was hardly a penniless garage shop. The CEO was a super-qualified MIT grad with rich parents back in India. All the engineers were top of the line and they worked out of prime technology real estate in the center of San Francisco. Their technology was leading edge for what you could do with digital radios and only one or two companies could come close to competing. They were financed by major venture firms both in Silicon Valley and on the East Coast.
A deal was struck between Vnet and Cliff and working together, we installed a complete Internet base station on Cliff’s signal hill, along with the familiar tall and narrow antennas. CPE (Customer Premises Equipment) Internet transceivers were distributed to willing customers and the system went live. Within a very short time, over 100 early adopter customers were online, growing steadily.
What we came to realize at this point was that the Vnet company was suffering from internal cognitive dissonance. What I will call their Eastern investors and our Western investors had very different ideas on how to proceed.
The Eastern group favored finding a major cellular customer, like maybe a Verison or AT&T, to place a large order even before the equipment had been field tested. In constructal terms, they wanted to access a long, fast distribution channel immediately, which they thought would bring the fastest payoff for themselves. What they failed to understand was the dynamics of those long fast conduits of commerce and how they actually work.
There are a number of constructal channels involved in the complete description of a major business like a telephony operator. A complete rendering with their interactions would be a major project. They include:
- Technology adoption
- Market intelligence
- Fiscal channels
- Political channels, external and internal
- Operational channels
For this brief example, we are going to look at three channels and compare the “early adopter”and “build it and they will come” strategies from a constructal viewpoint:
- Physical distribution
- Information distribution
One of the genius elements of cell carriers is their very limited physical distribution channel. They ship cell phone base stations to the vaults at the bottom of every cell tower. These require little maintenance and over their lifetime, each one generates predictable revenue from the volume of short and slow individual phone calls they aggregate.
Behind these base stations are large information channels called “backhaul” which aggregate the calls and send them to network switches which in turn send them to switches at the destination, reversing the process to the receiver of the call. Backhaul channels are long and fast, generally microwave or fiber optics. The larger the channel, the more efficient it becomes. Therefore, many companies simply rent bandwidth on long-haul “common carriers,” even operated or shared by their direct competitors.
It’s interesting to note that the backhaul channels are quite flexible in their virtual size. A small operator like Cliff might pay only a few hundred dollars a month to provision his 100-500 customers. But for a large operator, the cost of managing many small channels, or alternately, merging them with their existing long and fast network would exceed any possible revenue that might be generated.
To summarize this part of the constructal flow analysis, large operators were faced with integrating a small number of non-compatible short and slow channels into their large and fast information flow system, even for preliminary market testing. An additional problem is that the equipment needed for the Vnet base stations is in addition to the equipment needed to support the traditional cell phone business - an inefficient parallel flow structure.
However, Cliff only had to integrate his short and slow customer channels with any of a number of long and fast operators who are specifically adapted to efficiently manage their small flow volumes. As their system grew, they could morph and evolve their flexible vasculature as needed under changing circumstances.
The Technology Adoption Problem
The East coast investors believed that somehow a big company would adopt and deploy the technology faster than a band of small operators. (This is another subject for further analysis.)
Big companies are designed to maximize the efficiency of large, fast and long conduits of commerce. Introducing variation is both risky and expensive. Therefore, technology change managers need to be very, very certain of what they are committing the company to do. Trials are conservative, time consuming and small.
While the East coasters thought that 100-500 customers in the Cliff system was absurdly small, it was in fact more than the major companies were willing to take on in trial systems.
What if it worked?
The problem of small companies dealing with very much larger customers has bedeviled business for eons. In this case, think about what would have happened if Verizon, or whoever, had gone all in for the Vnet technology as the East Coast investors hoped.
Vnet would have been faced with the task of evolving from manufacturing a few hundred units to a million or more within a year to meet the demands of a large flow system that fanned out to multitudes of end user customers.
Equally problematic is the fact that Vnet’s early designs were certainly not evolved sufficiently to justify such large scale distribution in their current form. This is an example of how development and manufacturing channels have to be sufficiently evolved and working together in order to access an existing flow system that is very large and moving fast.
Technology integration is only one of the challenges facing a large conduit operator trying to evolve a new business model. Marketing, customer service and all the channels mentioned above also need to be considered in the trial phase.
Every Vnet Internet receiver needed to be delivered and set up in the customer’s home. For Cliff, this was not a problem. A couple of low-paid techs could keep up with growth and in the process encourage satisfied customers to recommend the system to their friends, (early adopter fanout), a very cheap form of marketing.
For the large cell operators, it was a different story. One of the huge advantages of the cellular telephone business is that the company never needs to come to your home. You buy a phone, get a contract and start talking. It seems beyond obvious, but cell phone companies do not have established channels for home visits.
In addition, think of the marketing problem for a cellular operator in rolling out this new Internet service. You need to roll it out to the largest possible area immediately. For example, your long and fast marketing conduits cover all of Los Angeles, not just a neighborhood. In order to only market to a small area, you would have to develop entirely new short and slow marketing channels to serve that area.
What were they thinking?
That this was not immediately obvious to the Eastern investors, who were betting large amounts of money, is almost incomprehensible. They prevailed and committed the company to the big-customer course. The large operators played with them for awhile, more to get operational experience than any intention of actually deploying Vnet technology. Then Vnet was gone.
The Cliff system in rural Oregon survived for several more years, earning a modest income until the hardware finally gave out,
Ultimately, Moore’s Law brought data capability to ordinary cell phones. The customers paid for new cell phones and it’s easy to see that the cost of new base stations, incorporating many advances beyond Internet data at much lower cost, was trivial to the operators. Hard wired cable TV systems offering Internet access have also continued to vascularize ever more rural markets.
Wireless Internet access via a dedicated receiver, like the Vnet system, is still more efficient than the cell networks. In areas where cable TV operators have not penetrated (they operate in constrained conduits of commerce as well), Internet access via modern versions of the Vnet system, operated by small, short and slow system operators are still much in use. Here in Panama, on the front of my house I have a dish with an Intel WiMax radio inside giving me excellent wireless Internet access from the signal hill atop a volcano some ten miles away.