This article includes a one-page preview that quickly summarizes the key ideas and provides an overview of how the concepts work in practice along with suggestions for further reading. Nowadays, virtually no companies innovate alone. Firms team up with a variety of partners, in a wide number of ways, to create new technologies, products, and services. But what is the best way to leverage the power of outsiders? To help executives answer that question, Pisano, of Harvard Business School, and Verganti, of Politecnico di Milano, developed a simple framework focused on two questions: Given your strategy, how open or closed should your network of collaborators be? And who should decide which problems to tackle and which solutions to adopt? There are four basic modes of collaboration, say the authors. An elite circle is a closed network with a hierarchical governance: One company selects the participants, defines the problem, and chooses the solution. For instance, Alessi, an Italian home-products company, invited 200 outside experts in postmodern architecture to contribute ideas for new home-product designs. An innovation mall is hierarchical but open: Anyone can post a problem or propose solutions in it, but the company posting the problem chooses the solution. An example is InnoCentive.com, an eBay-like site where companies post scientific challenges. An innovation community is open and decentralized: Anyone can propose problems, offer solutions, and decide which ideas to use - as happens in the open-source software community Linux. A consortium is a private group of participants that operate as equals and jointly select problems, decide how to conduct work, and choose solutions. IBM has set up a number of consortia with other companies to develop next-generation semiconductor technologies. No one approach is superior; each involves strategic trade-offs. When choosing among modes, firms must weigh their advantages and challenges, and assess which will work best with their strategy, capabilities, structure, and assets.