Co-authored with my husband, Ben Horowitz.
Before you start building any new solutions, it’s important to understand the difference between innovation and invention.
This is an important distinction because invention is more expensive than innovation. I use the word expensive, not about money but about all resources. It takes more time, money, effort, and opportunity cost to invent a solution than to find an innovative one.
Innovation is simple
Innovation is very straightforward. You take a solution that you know works in one place, and you apply it to a similar problem in a different space. Break down the problem you’re solving into its most simplistic form. What is the core of the problem? Then, look to other industries and disciplines to see if there were similar problems that have already been solved. You may be able to introduce an old solution in a new way. That’s innovation.
Invention is complex
Invention is a different mechanic for problem solving. Invention is where you start for problems that are totally unsolved. If you have no idea what might work, you start structuring experiments to chip away at the unknown. When you find a few things that work a little bit, you can expand on that knowledge until you find a solution (maybe, but not always.) Invention is a last resort when no other known solution will work.
If “chipping away at the unknown” is how you’d describe your backlog, then you might be inventing a solution. Try looking for an innovative solution first. What you find might not work perfectly, but it may significantly reduce the scope of the unknown. Then fewer experiments and feedback loops are needed to succeed.
Pattern matching on the problem
There’s an excellent example of the cost-saving benefits of innovation in this HRB article:
“One of the best innovation stories I’ve ever heard came to me from a senior executive at a leading tech firm. Apparently, his company had won a million-dollar contract to design a sensor that could detect pollutants at very small concentrations underwater. It was an unusually complex problem, so the firm set up a team of crack microchip designers, and they started putting their heads together.
About 45 minutes into their first working session, the marine biologist assigned to their team walked in with a bag of clams and set them on the table. Seeing the confused looks of the chip designers, he explained that clams can detect pollutants at just a few parts per million, and when that happens, they open their shells.
As it turned out, they didn’t really need a fancy chip to detect pollutants — just a simple one that could alert the system to clams opening their shells. “They saved $999,000 and ate the clams for dinner,” the executive told me.”http://bit.ly/2S3rOEK
Stanford bioengineers develop a hand-powered centrifuge, dubbed the “Paperfuge.” How might you achieve the functionality of an expensive medical machine in places with no infrastructure? Well, break the problem down to its simplest form: what’s cheap, doesn’t need power, and spins really, really fast?
“It is an ultra low cost centrifuge that’s built out of principles of a very old toy, the whirligig.
… We were able to essentially make a centrifuge that spins all the way to 120,000 RPM and 30,000 g forces. In the lab, we can separate and pull out malaria parasites from blood, we can separate filaria, African sleeping sickness, separate blood plasma.
… We can match centrifuges that cost all the way from a thousand to $5,000, but this is a tool that requires no electricity, no infrastructure. You can carry them around in your pockets for a price point of 20 cents.”http://bit.ly/3b12ZSA
My favorite TED Talk was given by Tal Golesworthy, an engineer with a heart condition who realized that weak arteries at risk of bursting look a whole lot like a standard plumbing problem. He outlines some of the unique challenges his cross-functional team encountered while designing an implant that externally braces the aorta, preventing rupture and negating the need for open heart surgery.
These stories are of cool, interesting, and successful ideas. But that’s not what makes them innovative. We already knew these solutions solved problems — detecting pollution in water, bracing leaky pipes, creating centrifugal force with only manual power. We understood the mechanism by which they solved these problems. The innovation was recognizing that a similar problem existed in a different space that might also be solved by these existing solutions.
Multidisciplinary teams and diversity
These examples have something else in common. Each of the teams responsible for these innovations was made up of people from diverse backgrounds. Someone with seemingly unrelated knowledge happened to pattern match the problem they were facing here with a problem they had seen before in a different domain, to arrive at solutions beyond the body of knowledge associated with that domain.
- The marine biologist solved the pollution sensor problem in a way that the chip designers would never have known about.
- Tal Golesworthy applied his knowledge of plumbing to his circulatory system in a way doctors didn’t know could work.
- The Paperfuge problem was solved when one of the team members remembered a childhood toy from India that spun really fast, and decided to analyze how it worked.
And, the similarities were immediately evident to the people who had seen these problems before. It may still have taken a while to finesse the solution to fit the new problem space, but that initial spark of innovation happened right away. Humans are very good at pattern matching — often too good, in fact. Pattern matching on the problem is where innovation begins. More diversity on your team means more opportunities for these happy coincidences to ignite innovative ideas.
If you’re trying to fix an innovation problem, start by fixing a diversity problem. Do you have a wide breadth of knowledge, education, and experience on your team? Homogeny inhibits innovation.
The high costs of failing to innovate
I once failed to turn an EEG device into a music input into an instrument by mapping it to MIDI sounds. EEG files were gigantic at the time, and being in my early 20s, and couldn’t afford enough hard drives for the project. So I was researching EEG file compression. I came across a paper written by PhD students from the University of Malta, who had spent their entire semester and grant trying to apply a JPEG compression algorithm to EEG files. My husband, the audiophile and engineer, was dumbfounded. EEG is wave data. We have existing — efficient! lossless! — compression algorithms for wave data: any kind of audio compression. Inventing a way to compress wave data with a lossy image algorithm struck him as absurd.
That’s why it’s important to investigate your innovation options before you start inventing. In this case, inventing a solution was wholly unnecessary, and in fact, the result has caused diagnostic issues. The lossy compression format can omit “diagnostically relevant” data for patients with epilepsy. With a more diverse team, this could have been avoided. A computer scientist or an audio engineer would have helped the neuroscientists select the right algorithm for the problem of compressing wave data.
Use innovative hypotheses to shorten your experimentation process
I am a huge proponent of experimentation and feedback loops — trying things, learning things, and feeding that knowledge back into the experimentation process. But it may not always be necessary to start from square one. Put together a diverse team. See if you can start with an informed hypothesis from solutions in other spaces. Do your research and amplify other’s efforts. That’s innovation.