Building an innovation engine is about how to set up a system for generating innovation, using a lean process that Scott Anthony, David Duncan, and Pontus Siren described in HBR.

Too often, companies don’t develop innovations in an orderly or reliable way. Big breakthroughs happen only because of one person’s heroic efforts or a huge dose of serendipity—or both.

Some firms try to spark innovation through ad hoc initiatives such as hack-a-thons, contests, and task forces. But these efforts often prove fruitless—or generate ideas that don’t fit the strategy.

A few companies, including P&G, go to the other extreme. They’ve built so-called innovation factories so they can attack the problem on a very large scale. But that requires a lot of time and money, new hires, and new organizational structure

There is a more practical middle ground: a minimum viable innovation system. It borrows from the lean start-up approach, which introduced the idea of creating a “minimum viable product”—a stripped-down prototype—to test new offerings with customers. A minimum viable innovation system includes only the essential building blocks for a reliable and strategically focused innovation function. You can set up your system by following four basic steps.