"If you're never making mistakes and never taking chances, you're probably not innovating disruptively."
-- John Cohn
IBM ( IBM 0.90% ) fellow John Cohn knows a thing or two about innovation. He has 116 patents and 36 technical papers to his name, and has helped direct IBM through four decades of computing innovation. From vacuum tubes to artificial intelligence (AI) and quantum computing, he's seen his share of technological change!
I recently spoke with John about several innovations on the horizon. Artificial intelligence and machine learning inference are quantum leaps that could forever change the computing world as we know it.
We then focused on how companies should best adapt to these types of massive market changes. Specifically, we discussed how astute leaders need to strike a healthy balance between placing short-term and long-term bets on research.
Cohn acknowledges the need for incremental improvements. Smaller innovations to existing product lines are important for companies to maintain their competitive position against others, or to provide the sales growth necessary to fund operations.
Here's what he said about incremental research in our conversation:
Just like there's continuous innovation and disruptive innovation, there's incremental research and then there's disruptive research. Incremental research, for example, we have people who are working on the science of AI and the core algorithms of AI. Our job is to continuously improve the throughput or the capacity and the size of models. We can do that in an engineering approach.
Swinging for the fences
But Cohn also stresses the importance of companies making larger bets and investing in disruptive innovation. Companies should encourage disruptive research that could potentially fix the industry's greatest pain points. These types of innovations aren't linked to existing products, and would often even compete against its current business.
Here's his first-hand perspective on disruptive innovation:
But sometimes you need breakthroughs. They're predictable breakthroughs, where the bottleneck is. You actually have to invest. You have to constantly manage your portfolio that you have some stuff that's near end, and intermediate, and further out.
In a past life, when I was in the chip group, I helped coordinate my division's relationship with researching. We were very conscious about making sure that we never pulled the rope too tight, so we're making research into just an extension of development. Clearly, the opposite is a problem. Many companies that no longer have industrial research found that if they didn't put enough linkage to the actual product lines, that research would kind of wander off and become irrelevant. I think we've managed that very well. We pay a lot of attention to making sure that we've got a pipeline that's going.
I think one of the things that's very interesting, going back to how those disruptions work, something that's very important to me, is the importance of giving time. You were talking about how when you do quarter to quarter, if you're not careful, you sort of choke the life, the play, out of engineering. If everything is incremental, you don't have the time to take chances, occasionally make mistakes. Because if you're never making mistakes, you're never taking chances, you're probably not innovating disruptively.
Cohn gives a real-life example of how IBM is investing in disruption by building out a quantum computing division:
If you look at something like quantum computing, we're investing in some of that that's a little further out. It's a couple years down the road, but setting the seeds for that and actually doing the core technology right now. It's something that we're taking a very long view on.
We've been talking about quantum for 15, 20 years -- about the theoretical underpinnings of reversible computing, etc. Only recently do we actually have the technological wherewithal to be able to build a real working system. That is, of course, accelerating.
You actually have to invest. You have to constantly manage your portfolio that you have some stuff that's near end and intermediate and further out.
The investing takeaway
As investors, we should respect the challenges that leadership teams must simultaneously manage.
There is a necessary balance between the time horizons. A good management team should be able to develop and articulate annual and three-year plans in order to address short-term expectations. But it should also have a close pulse on the needs of its customers, and be willing to allocate research funds to longer-term innovation.
Cohn describes that these longer-term innovations are much harder to achieve:
You can't schedule breakthroughs. When you find them, you have to work [for] them. You always have to have multiple horizons.
You have to make sure that you have some near-term stuff that's paying the bills, but some longer-term bets that you're actually getting in at the beginning, when the value is low and it's a time for growth.
It comes back to that risk tolerance. If you're always looking for the short-term win, you're always going to be on that incremental growth, right? If you've taken a couple of those really long bets and doing the research to figure out who's actually writing code and who's writing PowerPoint, I think that that can help you make a long-term portfolio.
As long-term investors, we should take special note of companies that are also committed to their own long-term research.