Kennametal Exec on Innovation Governance & ‘Fast Fail’ Testing

By Scott Kirsner |  December 22, 2014

Colin Tilzey heads the Innovation Ventures Group at Kennametal, a $3 billion supplier of tooling, wear components, and advanced materials, based in Latrobe, Pennsylvania. The company makes everything from drill bits to snowplow blades to earth-cutting consumables. Tilzey reports to the company’s Chief Technology Officer, with dotted line reporting to the Chief Marketing Officer. Innovation at the company is “seen as half technology and half business,” Tilzey says, and he sits in on both tech and marketing meetings.

Tilzey’s team consists of 18 people, about half Kennametal veterans and half new to the company. He says improving products and technologies in current Kennametal markets “is where 75-to-80 percent of our resources go. About five percent goes to completely new markets, and 15-to-20 percent adjacent. Ideally, we want it to be a 70/20/10 model.”

How our Innovation Governance Council has Evolved

“We have a governance council that consists of our CTO, CFO, two business unit owners, the CMO, and the global head of manufacturing. We meet with them every two months. We don’t tend to wait for permission to move ideas into incubation, we just ask forgiveness. Taking an idea from the PowerPoint and talking to customers stage to actually testing the idea can take anywhere from six months to three years. We report on that progress to the governance council.”

“Sometimes, we recommend killing something, and the governance council says, ‘Keep it going and explore a new avenue.’ Or we want to keep going and they want to kill it. Maybe there’s something happening elsewhere in the company, like a product line that we’re divesting. They may know about that while we don’t.”

“The council used to meet quarterly, and it lasted about two-and-a-half hours. We’d present an update on every project. But it was hard getting those executives all in the room at once. If more than two people weren’t going to be there, we’d cancel it. So we found we were force-feeding all these project updates to them. We shortened it and changed it to every two months, however we still cancel it if two people are going to miss it. We created two-page written executive updates for every project. Those get put into a pack, but we only do a deep dive on one major project at each meeting. Now, we have six meetings a year, and we shortened them to about one-and-a-half hours.”

“Without those meetings, and the input from the governance council, we wouldn’t get anything done. If I wanted to go spend $1 million, going through the corporate budgeting system, it wouldn’t happen. But if you have the CFO and CTO and business unit guys on board, when you go to the appropriation committee, it has already been approved. There’s no discussion of whether it’s a good idea or makes sense for the business.”

What’s the Cheapest Thing You Can Test?

“We call our approach the ‘fast fail’ model. We’ll try and test our hypothesis by piloting the cheapest success criteria in order to change direction, move forward, or decide to kill the project. We’re looking at killing 75-80 percent of projects. I get criticized if I’m not killing enough projects. So we’ll order rank all the things we need to test to make something happen, and cascade them as the cheapest, easiest, or quickest things to do. We figure out which is the cheapest part of an idea to test. Could I pay an intern to test something in the lab, or maybe adapt a machine we already have? Why not do that first, rather than spending a lot of money to do ‘voice of the customer’ research?”

“We kill most often at the ideation stage, or very early into incubation.”

“One of the measures we have is time-to-prototype, rather than time-to-market, or time-to-cash. And that can be applied to a business model, a mill, a drill, a piece of material.”

What We’re Working on Next

“Earlier this year, I created a commercial director position in my group who focuses on transitioning things back to the businesses. They want a crisp business plan, target customers, and materials for the sales guys. So that’s kind of a work-in-progress.”

“You want to have a mix of short-term ideas that you can give to the businesses fairly quickly. But you also need things that will have bigger impact. We’ve just done a system that logs tooling data in the cloud, and provides recommendations about how to improve productivity with a particular tool. We’ve beta tested it in our internal plants, and we’re just moving to external beta with customers. But we’ve never sold software before. We’re talking about what are the barriers to adoption, and how do we take those away? It’s turning into a real business, which is great. Our new CEO just came in in November, and he is meeting with us weekly. He is encouraging us now to go recruit more people, incubate it more. That is a good sign to us.”

More: Download several slides from Kennametal on innovation culture and governance in our Resource Center.