While counterintuitive, the reason why big innovations are truly big is that they invariably have something stupid about them that keep less innovative people from attempting to innovate. It’s far easier and less risky to innovate in small incremental steps. Ironically, people . . . the media in particular . . . having argued something was stupid, with 20:20 hindsight will then say the innovation was obvious and the inventor a visionary. What makes innovative people really innovative is their stubborn willingness to make the stupid work.
What’s so powerful about stupid innovations is that competitors rarely follow until it’s far too late. They fail to respond because the mental barriers erected from the seemingly stupid blocks their ability to see the value of the innovation. Their position is the losing one. So they will rationalize their position long past the time the evidence proves them wrong. It’s the ‘disposition effect’ applied to innovation. So, it’s important to embrace the stupid and risk failure because there can be huge payoffs.
Steve Jobs and Apple’s iPhone provide a clear example of this at work. When it was first leaked that this phone would not have buttons, pundits thought it was a stupid idea. How could you have a phone without buttons? A similar phone had been in R&D at Nokia well before Apple had theirs. Management killed the product because according to one involved engineer, “Management said, a phone without a keyboard will never sell.” Even after the iPhone was in the market, executives at competitors ignored it, hanging on the idea that the tactile feel of a keyboard was superior. They fully discounted the user interface value of programmable glass button images. On top of this, it was considered too expensive. Who would pay for it, when the big expansion of the market was for cheaper phones targeted at Chindia? Yet it was a huge success, disrupting the entire market. When Apple came out with iTunes, people thought that was stupid too. Who would pay for music when it was being passed around for free on the internet via Napster and others at the time? But Apple was responding to the way the world was changing as others rationalized free as being a point of no return.
Here are some of my favorite stupid innovations that went on to takeover entire markets:
ASML’s Twinscan: Two stages for a scanner? Insane! It will be too big, too costly, and too wasteful. But, Martin van den Brink was trying to figure out how to get to the next generation of wafer size (300mm) without taking a throughput hit. ASML had hit 100 200mm WPH and few believed this would ever be done of 300mm. Martin sat down, modeled it on a sheet of paper, figuring that with two stages and split tasks, enough dead time could be eliminated to pay for the cost and floorspace with more wafers out. Twinscan was a huge success. Not only that, the architecture paved the way for scanners that routinely get 200 300mm WPH.
Applied Materials’ Endura was dumb because it was hogged out of a single block of aluminum, which was way too costly. Dan Maydan stubbornly pushed it through because he knew leaks were the bane of all vacuum tools. He bet that the utilization and maintenance savings would offset the manufacturing cost of hogging out all that aluminum. He was right. Applied had never made a PVD tool before and yet its Endura would revolutionize the market as its competitors argued that it was too expensive.
Reliable semiconductor equipment was a stupid idea because semiconductor equipment worked at cutting-edge science and volumes were too few to ever make them as reliable as the Toyota some thought they should be. Chip equipment was more like a Ferrari than a Toyota. When SEMATECH started driving the industry to make tools more reliable mean time between assists was less than a 25 wafer lot and utilization ran around 25% of theoretical throughput; with 50 hour mean time between failure typical. But with dogged determination, SEMATECH in concert with SEMI provided the analytics, standards, and push that eventually made today’s tools arguably better than Toyotas.
HKMG was a dumb thing to work on because it was too difficult and process advantage could not be held. Process was supposed to come free with the tools you bought. Better to let someone else spend their R&D dollars on HKMG. IBM and Intel continued to work these problem technologies, knowing that if they did not, Moore’s Law would come to a halt. As a result, both companies and their allies pulled ahead. Intel was first to take Hi-K Metal-Gate to production, well over a node ahead of everyone else. IBM’s Common Platform partners beat their competitors by a year and took the majority share of the foundry market. Intel has parlayed its HKMG and now finFET advantage into mobile chip sets that are taking share in a market where it had no success on the basis of architecture alone.
Noyce’s IC was dumb because, theoretically, aluminum would not stick to glass (SiO2) and a metal was essential for interconnecting the transistors. Moreover, the law of large numbers meant it should not yield. No one attempted this at the Eastern electronics’ giants, because if they did and it failed, it would have killed their careers. Back in California, Gordon Moore walked into Noyce’s office and said, “Bob, we’ve tried everything but aluminum. What we tried won’t work and we know aluminum won’t stick to glass.” Noyce replied, “Why don’t you try aluminum.” The theories were wrong. Evaporated aluminum did stick to glass, could be patterned, and yields were not an issue. Noyce’s “Why don’t you try …” led to his patent on the IC (which, by the way, is the IC that would sweep the industry, as Kilby’s patent did not provide the economic advantage). Without his cavalier “Why don’t you try …” Moore’s Law would have never been possible and the semiconductor industry would not be near as large as it is today.