Business Model Evolution
During the last two decades, many of the old vertically integrated enterprise/empires have splintered into numerous core-competency-focused firms, loosely bound into the so-called virtual extended enterprises. The next two decades will witness re-integration into transenterprises – tightly knit federations of companies with long-term commitments and deep investments in integrating together into a more efficient and effective whole. The governance and investment models for these entities will transcend traditional strategic supplier relationships, combining the characteristics of vertically integrated companies and virtually integrated extended supply chains.
Transenterprises will go far beyond traditional strategic supplier relationship along two dimensions:
- Extraordinary commitments – financial and legal – massive investments and deep or exceptionally long contractual commitments binding the players into long-lasting transenterprises.
- Extraordinary integration – physical, process, IT systems, policies, and/or governance integration to a degree normally found only within a single enterprise.
Examples of Early Transenterprises
We are already seeing early versions of transenterprises. In the automotive industry, there is a movement towards supplier parks. In supplier parks, the automotive OEM selects a group of key suppliers to co-invest in a physically integrated factory. The suppliers’ manufacturing lines are housed on the same campus, generally in the same building, with their operations literally connected via conveyor belts. The net result is a large vertically integrated plant/campus with raw steel coming in one end, and finished autos out the other. This is physically very similar to the massive, vertically integrated automotive plants of the early twentieth century, but with a very different ownership and governance, because the assets are owned by many different firms. Suppliers invest hundreds of millions of dollars in equipment, and commit to decade long leases. The OEM takes on massive switching-cost risks. In some cases, suppliers have agreed to common HR governance across all the companies in the park to set compensation policies for the transenterprise, superceding the individual company policies.
For very different reasons, we see the genesis of transenterprises in the semiconductor industry. 25 years ago, many of the high tech OEMs made their own semiconductors. By the 90’s, most OEMs outsourced the design and production of custom chips. A complex web of players is required for this – “fabless” ASIC companies who design the chips, foundries who manufacture the “dies” (silicon, heart of the chip), assembly and test companies who package the chip, EDA companies that provide layout tools, others that provide the actual design building blocks, and companies who specialize in tweaking the manufacturing and testing processes to maximize yield.
Until about 3-4 years ago, OEMs could just “throw over the wall” their chip logic requirements to ASIC designers who would worry about the chip layout, who could then “throw it over the wall” to the foundry to worry about production and yields. However, as chips move to smaller and smaller geometries (with individual elements only a few atoms across) it has been forcing all the players in the chain to integrate their design, production, and test efforts much more tightly. They are being forced by physics to make significant investments and commitments earlier in the cycle of each new generation, in order to gain the yields necessary to be profitable. The design teams, from functional, to physical, to manufacturing, will need to integrate much more closely, much earlier in the overall cycle, and for longer durations. Companies will make longer-term commitments to specific platforms, architectures, and production processes with specific partners. With each new generation of semiconductors, these companies will come closer to building true transenterprises.
The emergence of transenterprises will look very different in different industries. Elements of transenterprises have existed in the chemicals industry for a while, such as fence plants–where a company builds a collocated plant to take the byproduct of another company’s plant, convert it, and in some cases sell it back to the first company. These represent very long-term commitments of assets. In automotive, the desire for build-to-order mass customization, the complexity of manufacturing integration, and the need for speed to avoid transportation delays are driving companies to co-locate. In semiconductors, the mind-boggling and ever-increasing challenges of moving to atomic and sub-atomic geometries are forcing the companies in the chain to integrate and invest together to create end-to-end solutions to solve these problems in a Moore’s law timeframe. The substantially different nature of transenterprises in automotive and high tech is illustrated in Table 1.
The Best of Vertical and Virtual
Transenterprises are not a return to the old style vertical integration, because the reasons for moving to the newer virtual integration in the first place, have not gone away. Companies are still shedding assets to improve ROA and agility. To increase their own competitiveness, companies are giving business units the freedom to go outside the company for shared services such as IT, and allowing buyers to purchase components from competitors, even though another division of the company makes that component – in essence, forcing internal functions/divisions to compete on the open market. This gets rid of bureaucratic fat and leads to a focus on core competencies, outsourcing things that someone else does better, and turning your internal excellence into business expansion. Transenterprises combine the best of vertically integrated firms (tight integration) and virtually extended enterprises (market competitiveness).
Another important point is that transenterprises will not be “old-boy” networks like keiretsu. Participation is entirely based on performance, best fit, and contribution, not simply having a long-standing relationship. Even though switching costs are higher than a traditional relationship, there is always the threat that you will be fired.
In industries where transenterprises are emerging, suppliers must look far ahead to understand critical criteria their customers and partners will use in selecting members of the transenterprise. OEMs will need the vision, boldness, and exceptional discernment and skill to select the partners to create winning transenterprises. Successful firms will be prepared to make the extraordinary commitments and take the extraordinary risks to be players in these winning entities, which will come to dominate their industry sectors.
A) Longer life and increasing cost of underlying processes– 20 years ago, a state-of-the-art fab (fabrication facility for chips) cost a couple hundred thousand dollars. Today it’s over $2 Billion. Five years ago, a mask (required for each iteration of each design) cost about $50K. Today it’s $1M. This trend creates strong pressure on semiconductor firms to increase the useful life of each fab and each design to get the ROI.
B) Shorter life and decreasing cost of high tech products – 20 years ago, the product cycle of a computer system was about five years. Today, it’s about five months.