Good organisation design can help companies become more innovative, but first you have to understand the nature of “new ideas” and how to generate them.

Not long ago, I was teaching organisation design to a group of senior executives when a somewhat tangential but insistent question rang out: “What can you tell us about innovation?” For today’s business leaders, it’s an understandable obsession. Most agree that the need to innovate is no longer confined to Silicon Valley; rapidly changing business conditions force a heavy and constant imaginative burden on all businesses that want to remain in the game long-term. There is no dearth of consultants and advisors offering to help “make your organisation more innovative” (for a fee). What should you know before you buy?  I have listed here what I shared with the executives during class: some of the most reliable old ideas about new ideas that I know of.

  1. Constraints are good for innovation

Most innovations aren’t necessarily the product of companies throwing large amounts of money at a problem – quite the opposite. New ideas have historically arisen when individuals were confronted with a problem they couldn’t buy their way out of, and so had to experiment within constraints. Constraints force us to discard preconceptions about processes and products. They require companies to rediscover their essential goals and improvise new, streamlined ways to achieve them. Be sceptical of the prevailing stereotype that innovation is about blue-skies thinking with lots of resources at your disposal. Creativity, ingenuity, and finding quick fixes and workarounds are all compatible with innovative thinking. By implication, so are organisational attributes such as performance targets, accountability and cost discipline.

  1. Most new ideas are re-combinations of old ideas.

Case in point: the iPod, undoubtedly an iconic innovation in recent memory. It left the global music industry forever changed, yet from a technological standpoint, it offered absolutely nothing new when it emerged in 2001. Already existing mp3 players contained all its functionality and more. New ways of combining old ideas are at the heart of innovation. Hence the continued importance of innovation hubs, from the coffee shop where intellectuals gathered in 18th century London to Silicon Valley, where spontaneous interactions among imaginative people have produced miraculous results. That’s also why I stand with Yahoo! CEO Marissa Mayer in opposition to the idea that employees can be as creative in isolation at home as they would be working with colleagues in the same physical space. There is an important connection between creativity and proximity. Design the organisation in a way that allows “ideas to have sex” as British author Matt Ridley puts it.

  1. Most new ideas are bad ideas

Organisation theorist Jim March taught us about the exploration-exploitation trade-off: If we keep re-using what works, we will never discover anything better; but the search for something better may well turn up duds. The implication: it’s not crazy to throw some money at crazy ideas. Example: In 2001, McDonald’s in Switzerland opened a pair of four-star branded hotels that lasted just two and a half years. While on the surface a baffling move, the hotel project was in fact a direct response to a call from HQ for experiments and ideas to help expand the brand. The fast food giant was aware even then that diversification was needed to offset the long-term threat that increasing global health-consciousness posed to their burgers-and-fries business model. The “Golden Arch” hotels certainly lost money, but the loss was no doubt deemed acceptable, incurred as it was while the company was flush with cash, and sacrificed as it was at the altar of the prospect of innovation. If you want to innovate, learn first to tolerate failure.

  1. New ideas are necessary, but not sufficient for innovation

Producing a good new idea doesn’t mean it will make you money. The classic illustration of this principle is the story of Xerox PARC, an R&D division that in 1973 produced the Xerox Alto, the world’s first personal computer and a major influence on all the commercial PCs that would follow. It was also responsible for the first prototypes of the laser printer, graphical user interface (GUI) mouse, and other technologies that were to become staples of home computing. Yet, in the end, Xerox failed to capitalise on its legendary innovation capacity, and PARC could only sit back and watch as Steve Jobs and others reaped extremely lucrative inspiration from its inventions. Myriad theories have sprung up to explain the failure of Xerox to capitalise on PARC’s new ideas. I don’t have the space to go into them here – but the case suffices to illustrate why having new ideas is not the same as being able to monetise them.

  1. Organising for generating new ideas is not the same as organising to use new ideas effectively.

The McDonald’s example above demonstrates that generating good new ideas requires a certain amount of tolerance for failure, but clearly this can’t be the standard for parts of the organisation that must specialise in implementing good ideas.Manufacturing processes, by necessity, are considered successful only if they are able to provide flawless replicability and reliability, something that would be considered an insane expectation in an R&D department. Organisations must take these differences into account either by making room for both kinds of contexts in different parts of the organisation, or by abandoning the dichotomy altogether in favour of specialisation. The pharma industry has seen a sharp division of labor between risky idea generation by biotech firms and commercialisation by big pharma.

Phanish Puranam is the Roland Berger Chair Professor of Strategy & Organisation Design at INSEAD.