Open innovation is not outsourcing

The arithmetic behind open innovation is uncomfortable for any research director. However large the laboratory, however generous the budget, nearly all of the competent people in any field work for someone else. A firm that learns only from its own scientists has decided, in effect, to ignore most of what the world knows.

For most of the twentieth century that decision looked reasonable. The great corporate laboratories, Bell Labs, Xerox PARC, the DuPont Experimental Station, were built on a closed logic: fund the research, own the results, protect them with patents, and carry them to market through your own products. Knowledge was scarce, concentrated and slow to move, so vertical integration of the whole journey from discovery to distribution was not arrogance. It was the efficient design for its time.

Henry Chesbrough's contribution, in 2003, was to name what had already quietly broken. Skilled researchers had become mobile, carrying knowledge out of the building with every job change. Venture capital gave them somewhere to take it. Universities had grown from teaching institutions into serious sources of commercial science. Product lives had shortened, which meant a discovery hoarded was increasingly a discovery wasted. Knowledge stopped respecting company walls, and the closed model stopped being efficient and started being expensive.

Open innovation, in Chesbrough's definition, is the purposive use of inflows and outflows of knowledge: external ideas brought in to accelerate internal innovation, and internal ideas released to markets the firm will not serve itself. Both halves matter. The second is almost always forgotten.

Inbound open innovation is the familiar half. Firms license technology rather than reinventing it, station scouting units near universities and startup clusters, develop with suppliers and customers rather than merely buying from one and selling to the other, and acquire small companies as much for their teams as for their products. Eric von Hippel's work on lead users showed how much of this potential sits in plain sight: in field after field, the most demanding users modify products long before manufacturers notice the need, which makes them a research department nobody is paying for.

Outbound open innovation is the neglected half. Most large firms commercialise only a fraction of the patents they hold; the remainder sit as dormant assets, retained mainly to block rivals. Licensing them out, spinning out teams whose projects no longer fit the strategy, or selling intellectual property outright converts shelf inventory into revenue and, sometimes more valuably, into influence over where an industry's standards settle. A technology released into the world keeps working for its originator in ways a technology locked in a drawer never can.

Between the two sits coupled innovation: alliances, co-development agreements, platforms and open challenges, where knowledge moves in both directions under negotiated rules. The celebrated examples are easy to admire and easy to misread. LEGO Ideas has drawn hundreds of thousands of members and well over a hundred thousand product submissions, of which a small fraction have ever reached production. NASA's open challenges have produced answers to problems internal teams had carried for years. What makes these programmes work is not the size of the crowd but the quality of the filter: a precisely framed question, a credible evaluation process, and a clear path from selected idea to shipped result. The crowd supplies variance. The institution must supply selection.

Outsourcing buys a known output at a lower cost. The task is specified first and contracted second; the relationship is transactional; the buyer's ambition is to stop doing something. Open innovation has the opposite ambition. It changes where the firm's questions get asked, not merely where its answers get produced, and done properly it raises the demands on internal capability rather than lowering them.

The reason was articulated by Wesley Cohen and Daniel Levinthal under the name absorptive capacity. A firm's ability to recognise the value of external knowledge, assimilate it and apply it commercially depends on the prior related knowledge it already holds, and that prior knowledge is built mainly by doing research, not by reading about it. A company that guts its laboratory and declares itself open soon discovers that it can no longer tell good external science from confident nonsense. It has outsourced not just the production of knowledge but judgement about knowledge, and judgement was the asset that mattered.

This is the paradox at the centre of the model: the more a firm intends to borrow intelligence, the smarter it must remain. Procter & Gamble's Connect and Develop programme, the canonical corporate case, set a public target for the share of new products with an external component and met it. What the retelling tends to omit is that P&G did not dismantle internal R&D to get there. It redirected it, from generating every idea towards also evaluating, absorbing and scaling ideas wherever they originated. The laboratory's job widened; it did not shrink.

Openness is not free, and its costs arrive in unfamiliar ledgers. Searching the outside world consumes attention, which is scarcer than budget. Keld Laursen and Ammon Salter found that the returns to external search rise and then fall: firms that scan too many sources too superficially perform worse than those that go deep with a chosen few. Over-search is a real failure mode, not a theoretical one, and it is the natural failure mode of a firm that has confused the posture of openness with the practice.

Intellectual property strategy changes character. In a closed world patents are fences. In an open one they are also currency: the instruments that make collaboration governable and trade in knowledge possible at all. Managing that shift requires legal capability of a different temperament from pure defence, comfortable writing licences as often as litigation threats.

The hardest costs are cultural. Not-invented-here bias leads teams to undervalue external work precisely because accepting it feels like admitting defeat. Its mirror image, sometimes called not-sold-here, kills outbound licensing because releasing a technology feels like arming a competitor even when the firm has no intention of ever using it. Both reflexes are rational at the level of the individual career and wasteful at the level of the firm, which is exactly what makes them durable.

The practice, as opposed to the posture, has a recognisable shape. The question comes before the channel: firms that decide what they need to learn and then choose the mechanism, a licence, a partnership, a challenge, an acquisition, do better than firms that launch a portal and wait. The funnel is instrumented: time from external submission to decision, share of the portfolio with an external component, licensing income, and the conversion rate from pilot to scale are all measurable, and what is not measured will quietly revert to closed. The partner portfolio is balanced deliberately, because different partners answer different questions: universities stretch the horizon, startups compress time, suppliers attack cost and quality, lead users expose the jobs the roadmap missed.

And somebody owns absorption. The recurring design mistake is to make external knowledge everyone's second job. The firms that profit from openness typically maintain a small, senior, technically credible team whose first job is to scout, evaluate and shepherd external work into the line organisation, with the authority to say no early and the standing to make yes stick.

None of this makes openness a creed. When the critical knowledge is tacit and architectural, spread across the seams of a complex system rather than contained in tidy modules, integration beats markets: some things must be developed under one roof to work at all. When the appropriability regime is weak, when patents are hard to enforce and imitation is fast, secrecy may be the only protection that holds, and secrecy and openness make poor housemates. When the bottleneck is not a shortage of ideas but an inability to integrate the ones already in hand, opening the front door merely lengthens the queue.

The honest position is contingency. Openness is a dial, not a switch, and its setting deserves the same rigour as a pricing or capacity decision: what must we know ourselves, what can we responsibly borrow, what should we let go.

The lasting shift in perspective is quieter than the slogans suggest. Open innovation reframes the firm from a container of knowledge into a node in flows of knowledge, and turns the corporate boundary from an inherited fact into an instrument of strategy. Firms that hear this as permission to spend less on knowledge get the worst of both worlds: they lose the internal depth that made external ideas legible in the first place. Firms that hear it as a discipline, a deliberate architecture of inflows and outflows maintained by people capable of judging what flows past, compound their own learning with everyone else's. The corporate laboratory is not obsolete. It has acquired a second job: not only to invent, but to make the firm worth trading knowledge with.