Many believe without question the idea that real innovation only occurs in small companies and that big companies are too stodgy and quarterly number-focused to innovate. Sure, some say big companies may innovate in an evolutionary way, but never with revolutionary ideas. While this may be true in other industries, the evidence that it is false in high-tech abounds.
It is true that many high-tech companies have and will fail because they stop innovating. Blackberry and Nokia are recent examples. But there are plenty of others who have systematically reinvented themselves over time like IBM, Apple, and Intel. IBM has a hundred years of doing it. What’s really at work here has little to do with size.
This is a really important question today, because the rising scale of innovation means that the semiconductor industry long is past its era of venture funded start-ups. Does that mean that innovation must stop? We see a shade of this question from those who argue that Moore’s Law has stopped or slowed for scale reasons.
According to Darwin, survival of the fittest does not mean that the strongest survive. Nor does it mean that the smartest survive. Those that survive are the ones most adaptable to change. Companies that don’t adapt are soon relegated to history’s dust bin.
With respect to companies, think of the above and replace strongest with biggest and smartest with smallest. Big companies are strong because they have the financial muscle to do what they want. Small companies are smart because their light management structures mean they can execute decisions faster. But in reality, both fail when disruptive technology comes along and they fail to adapt. Larry Page puts it well when he says, “Big companies fail when they fail to see the future.” But this is true of both big and small companies.
Change, as I first wrote in Maxims of Tech, is the one thing that makes the management of a technology company so much more difficult than traditional companies, which only have to deal with labor, infrastructure, and capital issues. It’s why business giants like Warren Buffet and Jack Welch have said they steer a clear path of technology companies.
Change is innovation. Innovation is change in high-tech. Innovation is what drives sales in high-tech. Sales stagnate and companies falter when they fail to innovate. Stepping back from the strict definition of Moore’s Law, it is clear that what Gordon saw when he first penned his paper in 1965, was the result of the relentless exponential rate of innovation. As the world turned to science it began to answer many questions. For every question it answered, it raised at least two more. As knowledge grew, what we knew about what we didn’t know grew more expansive. That exponential growth created the opportunities that made technology a big business in the twentieth century. But it also got us to a level where only large scale can address the problems we face.
So what is it about innovative big companies that allow them to innovate? Let’s look at a few and the principles that can be derived from them:
Apple: If ever there was an example that proved big companies can innovate, it’s Apple under Steve Jobs. Apple is a sterling example that size doesn’t matter because it’s gone through periods of great and failing innovation both as a small and big company. Moreover, under Jobs, Apple’s innovation process was tightly controlled and very dictatorial, which is supposedly a no-no. However, it gave them the advantage of fast decision execution times. Apple’s history proves that CEOs with great vision make all the difference in the world and dictatorial development processes work well under them.
IBM has reinvented itself many times over its hundred year history. In fact it was Lou Gerstner who popularized calling big companies ‘elephants.’ Unlike Jobs, Gerstner was not known for his vision and was even criticized for a lack of it when he started. But IBM didn’t need vision at the time. It needed focus and direction and that’s what he brought. It is virtually impossible for a diverse company like IBM to be led by the vision of a single individual. Gerstner found that IBM had plenty of people with vision similar to Steve Jobs. It was simply a matter of identification, delegation, focus, and alignment to overall objectives. That’s why IBM could break new ground in areas as diverse as copper interconnect, multi-core processors, and services. The most recent result is big data and architectures which led to Watson — a system that thinks on levels which artificial intelligence never dreamed of getting to.
Google is now a big company and Google Maps is an interesting example of faith in URL development model (Ubiquity first, Revenues Later). The URL model was how Google got started. They had no idea of how to monetize maps. At the same time, they knew it would require huge scale. They bet on the fact that maps are universally needed and that if they got people to use Internet maps for free, they could eventually figure out how to make money. It did, as smartphones evolved to make them essential and geo-based advertising models came into being with them. Moreover, it is maps that have made driverless cars possible. While not yet monetized, everyone now recognizes they are the future. The lesson is that the future starts at the intersection of how people use existing technology today and the potential of new technologies to change that. A second lesson is that there has to be an intense willingness to fail and learn as well has having what Larry Page calls, “A healthy disregard for the impossible.”
Intel: If you just looked at them, you would find every innovation principal from above. You don’t clock Moore’s Law for 45 years straight without mastering them. What is different about Intel is that they systematically drive their innovation inputs via emergent behavior across the supply chain. They cannot move forward with Moore’s Law unless they can get the tools and materials needed to do so on time. There are hundreds of suppliers, universities, and consortia that feed inputs in. So they must drive innovation via an emergent behavioral model. Most companies don’t or won’t develop their supply chain because their competitors use the same infrastructure. It’s hard to derive gains out of your supply chain. But remember that in the business of technology, the primary variable is change. Helping develop the supply chain derives efficiency gains from partnering and from being first with the knowledge of how to use a new technology. The time gained from getting to a new technology first is additive across nodes. So if you are consistently half a node ahead, you’ll be a full generation ahead after four nodes. So the lessons are, understand and drive the emergent behavior of your supply chain and go long for the gains.
Looking down into the supply chain you see big tech companies innovate not because they can, but because they must.
Synopsys is a dominant company in EDA. Their revenue model is fundamentally limited by the number of designers using their tools. Since transistor counts rise with Moore’s Law, a static model would indicate that their revenues would grow much faster if they didn’t innovate. Yet they continue to innovate in ways that increase designer productivity because the dynamic model is a tyrant for innovation. Without increases in productivity, design costs would soar, thus slowing down new designs and ultimately limiting the number of new end devices coming to market. People buy new smartphones and other devices not because they need a replacement; but because the new obsoletes the old. Thus, the EDA ecosystem is only kept healthy by improving designer productivity.
ASML is a dominant company in lithography. Their revenue model is fundamentally driven by Moore’s Law and they must innovate to keep it on track for the whole industry. If they don’t, Nikon will be right on their heels trying to take over. Both are big companies that systematically innovate. At the same time, ASML has driven huge gains in wafer throughputs which, with a static model, diminish demand for the number of tools they sell. If they don’t do this, exposure cost-per-transistor would not follow Moore’s Law down. If it went up, there would be little reason to scale; hence ending demand for new lithography tools; because without scaling, there would be fewer new devices, limiting demand.
Applied Materials is a dominant company in multiple wafer fab equipment segments. Their revenue model is fundamentally limited by the number of wafers that pass through their tools. Because of Moore’s Law and slowing semiconductor revenue growth the number of wafers made each year has hardly grown as well. A static model would be to cut R&D cost in exchange for immediate profit. However, dynamically, if they didn’t innovate, there would be little reason to buy new tools and upgrades from them. Without innovation, their revenues would decline. Moreover, cutting R&D for short-term gain is often a fatal mistake. It’s hard to get the people back and even harder to rebuild an organizational innovation engine.
Teradyne has long been a dominant company in ATE and they are the oldest. Their revenue model has been challenged for decades by a declining portion of semiconductor revenues spent in this area. Yet, as customers scaled down spending, Teradyne scaled down its organization and its testers. Innovation in organization is extremely difficult for big companies. Moreover, it fundamentally changes the message. It used to be that when you toured Teradyne you were impressed by its opulence — both literal and technical. The message to customers was typical of the nineties: if you want to be a king, you had better buy from a king. Now when you tour you are impressed by the depth of their cost consciousness. The message to customers is: if you want to control your costs, you had better buy from someone who is cost conscious. Scaling down its testers to levels like today’s respected FLEX series must have been hard for engineers used to designing technically complex ‘big-iron’ testers. But they did it because the market demanded it. The lesson here is that what the market demands, the market gets.
BESI is a packaging company that I have watched innovate in amazing ways over the decades in both small and large for its segments. It’s had a lot of tough competition over the years that have failed because they chose to stop innovating. Yes, BESI has a visionary CEO that always sees the future clearly. But most important he lives in mind and soul the challenge all technology companies must face of tough competition. If they don’t have tough competition they will if they don’t innovate. Dominant does not mean permanent. Innovation that addresses customer need is the only way to ensure permanency.
Key principles behind successful innovation in big companies:
· CEOs with great vision make all the difference in the world and dictatorial development processes work well under them.
· CEOs that lack vision must delegate it by empowering independent operating units with CEO-equivalents who can see the future. It’s about identification, delegation, focus, and alignment to overall objectives.
· Focus, direction, and alignment are critical to successful innovation.
· The future starts at the intersection of how people use existing technology today and the potential of new technologies to change that.
· There has to be an intense willingness to fail and learn as well has having what Larry Page calls, “A healthy disregard for the impossible.”
· Understand and drive the emergent behavior of your supply chain and go long for the gains.
· Big tech companies innovate not because they can, but because they must.
· Change drives dynamic models that are tyrants demanding innovation.
· What the market demands, the market gets.
· Dominant does not mean permanent. Innovation that addresses customer need is the only way to ensure permanency.
