In the field of innovation, the distinction between exploration and exploitation is universal. It is clear, it seems obvious, and it has become gospel in the world of innovation. Yet it is counter-productive, as it rests on questionable assumptions. It illustrates how the way we formulate a problem, i.e. our mental model, determines our ability to solve it. The wrong mental model locks us in, while the right one opens up possibilities. It’s time to let go the exploration/exploitation distinction.
The distinction was originally developed in a famous article by researcher James March in 1991. March writes: “A central concern in studies of adaptive processes is the relationship between the exploration of new possibilities and the exploitation of old certainties. Exploration includes such things as search, variation, risk-taking, experimentation, play, flexibility, discovery, innovation. Exploitation includes such things as improvement, choice, production, efficiency, selection, implementation, execution”. He adds: “Maintaining an appropriate balance between exploration and exploitation is a critical factor in the survival and prosperity of systems.”
This distinction explains the creation by large companies of entities dedicated to exploration: labs, creativity rooms, innovation departments, etc. And yet it reflects several highly questionable mental models.
Mental model #1 assumes that exploration is innovation. Here the risk is to confuse invention and innovation. We explore a lot of things but we don’t put anything on the market, or only half of it (Kodak). Exploration is the bane of large companies. They give themselves the illusion of innovating by doing “proofs of concept”, experiments or by multiplying patents. Nokia had projects of touch screens and smartphones long before Apple, but did nothing with them.
Mental model #2 assumes that innovation is exploration. An explorer is someone who ventures into distant lands to discover something completely new. Behind the notion of exploration is the idea of a significant distance in terms of knowledge between the current activity and the object of the innovation. However, this is not necessarily the case. Some disruptions can be cognitively very close to the current activity. EasyJet is disruptive for Air France, even though it is identical in terms of knowledge. There is no exploration delta between EasyJet and Air France. There are different choices in terms of business model, and that’s where the innovation lies. So innovation is not necessarily significant knowledge creation, let alone knowledge discovery.
Mental model #3 assumes that exploitation does not produce knowledge or innovation. Here the mental model underlying the exploitation-exploitation distinction appears in full, separating the noble domain, that of thought, and the inferior domain, that of execution. Only the former would produce knowledge. But what we call “execution”, more generally the world of operations, is extremely complex, and requires, and in return produces, a great deal of knowledge. The Japanese industrial power was thus built in the 1970s on micro-improvements drawn from the daily work in factories. In other words, the great Japanese manufacturing breakthrough did not come from exploration, but from thirty years of intelligent “exploitation”.
Mental model #4 assumes that exploration will solve the innovation dilemma. The dilemma is not the result of a cognitively distant activity, a significant difference in knowledge, but a conflict between the current activity and the new activity. Kodak’s problem in 1995 was that the new digital technology made its existing business (film) obsolete, and it was therefore reluctant to push it. Kodak will spend billions on digital, to no avail. It was neither completely exploration, since many products were launched, nor completely exploitation, since the field was very new. In other words, the exploration/exploitation distinction is not relevant to describe the innovation situation that Kodak was facing in 1995 and which is that of most of large companies facing a disruption in their environment.
Mental model #5 assumes that we can distinguish the nature of the projects a priori (exploitative, exploratory). We could thus put them in dedicated units, and each one would generate its own projects; but only the path of the project will orient it towards a more or less disruptive nature. Most disruptions are born from the improvement of an existing activity that at first seemed purely incremental but became disruptive. When we start, we cannot always know whether what we are doing will lead to an incremental innovation or initiate a disruption. This is the problem with companies that create entities dedicated to exploration. Many of these entities quickly become irrelevant because they work on white elephants that are totally detached from the organization’s activity. At the extreme, they are a playground for unmanageable young people or idle executives. Reserving disruptive innovation in an exploration entity is to take the risk that no disruption will ever happen, and that no innovation will be born in the exploitation part.
Mental model #6 assumes that exploration is relevant in uncertainty. Effectuation, the logic of action of entrepreneurs, shows on the contrary that in uncertainty, one must act. There is nothing to explore, one must create. When the two founders of Airbnb inflate an air mattress to allow someone to sleep in their living room, are they exploring? We’ve all done it to welcome a visitor to our home. It’s both a very mundane thing and the starting point for Airbnb to embark on a deeply disruptive entrepreneurial journey. It’s not exploration or exploitation; it’s creative action.
Don’t explore, create
The distinction between exploration and exploitation, while logical and appealing, is ill-founded and, above all, unhelpful, even counterproductive. If you are an entrepreneur or intrapreneur, what matters to you is not to create new knowledge, but a new activity. Don’t explore, create.