When Rene Descartes first published his meditations on first philosophy, they weren’t well received or acknowledged. It wasn’t until centuries later that the general community of philosophers found his work relevant and compelling. Today, he’s celebrated as one of the most important philosophers in history; however, he was never aware of his own impact because he was far ahead of his time. I see this happening with data science: The world isn’t ready yet for some of the data science that’s being developed.

Successful innovation with data science goes through a very tough cycle: a stroke of brilliance, a talented data science team that solves tough problems, and finally a market that gets it. If you strike out on this last step, you lose.

Weird science

Data science can make people aware of something they don’t want to know. For instance, Facebook recently raised a few eyebrows when its data science group was outed for manipulating the behavior of 700,000 users through news feeds. To data scientists, it may seem perfectly normal to mine through digital behavior to understand and ultimately influence future behavior. Marketing groups have been formally and publicly influencing behavior for decades, so why are Facebook’s data scientists any different? Because it’s data science, and that freaks people out. People want to keep scary science fiction to the silver screen until they’re ready to embrace it into their realty, which may be a while or never.

Reigning in your idea

You must work closely with your marketing department on your data science ideas. A good marketing department is worth its weight in gold when it comes to market sensitivity around product adoption. Products and services that involve data science are particularly risky in this area, so pay close attention to what they’re saying.

Out here in Silicon Valley, we’re notorious for letting product engineers run marketing, which is a very bad idea. It’s almost impossible for your product engineers, and definitely your data scientists, to see your analytic offerings through the customer’s eyes.

If you have a particular idea that may be risky for market adoption, it’s best to scale it into the market. As much as it might hurt your ego to bridle your brilliance, get your foot in the door with a more palatable version of your idea.

Imagine you’ve created an artificial intelligence unit for an automobile that’s similar to KITT — the fictional car that David Hasselhoff drove in the 1980s TV series Knight Rider. As much as I enjoyed watching Knight Rider as a kid, I don’t want my car knowing that much about me and what I’m doing. So even though this car could probably tell your customer that his brown shoes don’t go with his blue suit, that’s not a feature to make public just yet. It’s better if your car plays dumb a bit while still staying competitive, like automatically displaying traffic conditions and the best route to take when it knows he’s going to work.


Innovation with data science is exciting, but it can be risky if your market isn’t ready for your next great idea. Work closely with your marketing department to understand not only if, but when your next brilliant analytic offering will be a big hit. If it’s today, then go for it! If it’s sometime in the future, maybe you’re better off producing a science-fiction flick.