Dan Goldin | Crain's Philadelphia

In this ongoing series, we ask executives, entrepreneurs and business leaders about mistakes that have shaped their business philosophy.

Dan Goldin

Background:  

Founded by NASA veteran Dan Goldin, San Diego-based KnuEdge builds systems for machine intelligence and develops technology based on computational neuroscience. KnuEdge's products include military-grade speech recognition and authentication software.

The Mistake:

Early in my career, I thought that if we had a tough problem and we put enough people on it, we could get it solved faster.

I learned very painfully that the more people you have confronting a problem of innovation or of bringing a product to market, the more confusion you have. Suddenly, you’ve lost time.

Those five people ... put out the greatest speech recognition algorithm in the world.

The Lesson:

There wasn’t an aha moment, but after some decades in the business, I saw that there were too many people involved in each project. So I started experimenting by using fewer and fewer people. Over time, I saw that you could keep pulling people off a project without losing efficiency. It’s counterintuitive, but I found that the more people you pulled off, the more progress you made.

At KnuEdge, the whole innovation team is only five people. The assignment I gave them was: “I want you to be able to authenticate a person in the presence of noise, consisting of random noise like wind sounds, and structured noise like people talking in the background in a cafe, and I want that noise to be equal in energy to the energy in the voice. I want you to be able to authenticate the voice with very low error so there’s no false positive – like someone else pretending to be you.”

With five people, we’ve been able to innovate speech recognition and authentication software and have produced a computer chip that is industry leading. Those five people – two neuroscientists, one mathematician, one computer scientist and one project leader – put out the greatest speech recognition algorithm in the world.

I went from a staff of hundreds of thousands to a company of 100 employees. If you don’t have the exact skills inside your company, you find it outside and you outsource it. That’s how we’ve been able to keep a tight team and make a lot of progress.

It’s not that five is the magic number; there is no magic equation. You have to size up the situation and sort it through. The key point to keep in mind is that you need people who are self-confident, have a capacity for self-learning so that they don’t have to call in a bunch of experts, have a clear understanding of what the vision is for what needs to be accomplished, have a passion for their work, and seek excellence without compromise. You also need to select people who have a capacity to work with each other. You can’t select people who are brilliant but don’t know how to get along with others.

Follow KnuEdge on Twitter at @KnuEdge.

Do you have a good story you’d like to share, or know someone we should feature? Email cberman@crain.com.

And be sure to sign up for your local newsletter from Crain's.