James March has unique vision of innovation and he does not seem to take himself very seriously when it comes to innovation. But he is very serious.
Within a scientific approach, all theories are true until proven wrong, but above all, all current theories will one day be wrong in innovation.
If you do not believe in models and paradigms, and you believe innovation happens on the ground and not in academic papers, James may be your friend.
Change and innovation in organizations
James March does not have an idealistic vision of innovation. Most ideas are bad, maybe worse, as John Calvin once said, change is usually dangerous.
Lot of organisations invent, but few innovate as few encounter market success for their inventions.
Most organizations will produce solutions which are seeking problems
For James, innovations are successful only if it works on the long term. To him, innovators blame people that are not adopting the innovation fast enough but on the other hand, sceptics mostly think the innovation is useless, i.e. it does not solve a problem.
Garbage can model
The model name speaks for itself : most inventions will end up in the garbage can because most ideas are bad.
Over encouraging imagination can destroy an organisation because bad ideas outnumber good ones.
James March would not consider innovation as a real phenomenon, but rather focus on environnements in which decisions are taken. He has a backward approach to innovation, by going at the very beginning of it : how decisions are made.
We do not need innovation, we want innovation
“Innovation equals a bright future” they said
Most see INNOVATION as an escape from our present but going astray is usually rewarded with blood and tears in business. Consultant manage to “sell” innovation for 2 reasons : both a promise of survival for organizations and to avoid boredom amongst employees.
James warns us : innovation has to be a problem resolution process only.
Furthermore, innovation has to be a constrained process, with strict timelines and clearly defined objectives.
Balance between killing innovation and bankruptcy
A successful organization will pick good ideas, get rid of bad ones really quickly without preventing a tidy and rational ongoing business.
James thinks this balance is very hard to find in practice, that is why innovation rarely happens. Innovation needs plumbing (rationality) and poetry (irrationality).
James talks about “slacks” in organization, allowing room for individuals to take initiative. Slacks are surplus (time, resources,..) that can be dedicated to new ideas without jeopardizing current affairs.
According to March, a too efficient organization will kill innovation, including a too “Lean” company. Slacks are inefficiencies within the organizations. To James, these margins are good as long asthey are not dangerous in good times and they act as safety net in bad times.
Slacks are not conscious for managenent. Slacks are a delta between potential efficiency and current efficiency within an organization. As a consequence, slacks are room for improvement.
For March, organizations are systems which are structured by its many interactions. Organizations are not top-down structures.
Through danish philosopher Kierkegaard, March draws a parralel between religions and organizations, especially the role played by faith. Proving or trying to prove utility of an organization is the beginning of the end of this organization. Each organization needs faith.
Building on polity, where symbols and rituals create the structure, business are no different to March, and all organizations can be seen as a church. This is why culture matters so much for startup, as any new church seeking worshippers.
Innovation is about faith and can be translated as a value of eternal change for employees. Rituals have to be created around these values to structure the organization.
How to innovate in an homogeneous organization?
An organization is built around values, routines and past dependency with top-down objectives (strategy). It seems quite difficult to allow innovations within these constraints. James March thinks of change as truly endogenous and not coming from an exogenous paradigm shift.
To March, organizations are much more complex than the homogenous image they want to convey. This complexity lies in its ambiguities.
There are four main ambiguities : role constraints (the link between belief and action is filtered by one’s role within the organization); superstition (no rational link between environment inputs and action taken); audience (learning happens but past dependency is too strong for change to happen); and ambiguity itself (bad comprehension of problems where cause to effect is not clearly defined).
March thinks innovation and change come from people interactions within the organization. Trust within the organization is really important for people to interact.
What is ambidextry and how to set it up ?
March fundamental contribution is about ambidextry : how to allow both exploitation (short term amelioration through optimization and prediction) and exploration (innovation for long-term perspectives through experimentation, research, discovery and games…).
Exploitation is current knowledge improvement and exploration is new knowledge discovery
According to Duncan, this tension between exploitation and exploration is required in the long term but it is always a fragile balance. March underlines that too much focus on exploitation will jeopardize the organization in the long term (running of ideas) and too much strain on exploration will kill the organization in the short term (by running out of cash).
To set up ambidextry, there are three main options : structural (one entity for exploitation and one for exploration, vision supported by Schumpeter and Porter); contextual (depending one the moment, the team will exploit or explore, with seems to be the startup mode, [this vision is supported by Gibson-Birkinshaw]); and finally there is a network option (by exploring outside the organization through crowdsourcing and partnerships with startups).
The main issue with exploration is that, most of the time, there is no practical outcome, unlike exploitation.
March’s view can be understood as a criticism of “lean management” which may be seen as an obsession to refine current knowledge. March warns us about knowledge lock-ins, which may cause serious damage to the organization in the long term.
On the other hand, lock-ins may also emerge in exploring entities : if there are many failures in a row, agitation will create a resources consumption spiral; and if one success happens, this success will create a path-dependency for future exploration.
As a conclusion, ambidextry dilemna must be solved by itself pursuant to March, by letting this tension live in the what he calls “ecological exchanges” (within organization members and between organization and its environment).
For each organization, a evolving code of conduct
Each organization has a code of conduct, i.e. all the implicit and explicit rules according to which one member has to behave.
This code is constantly interpreted by members, and their resulting actions make this code evolve all the time. A lot of learning comes from this interpretation of the code and tensions resulting from the code. Tensions are necessary for change to happen.
Conclusion : a constructivist view of innovation
Agreeing with Weick model of enactment-retention-selection model, March has an action-focused vision of innovation, where what has to be kept can only be defined ex-post. Action has to be priviledgied over prediction for an organization to last.
This constructivist vision, where experiment is key, seems to more prolific than a positivist view, where control and prediction appear useless in such an uncertain world.
Because of bounded rationality, most organizations are irrational nevertheless they are willingfully rational, because rational actions are usually not taken mainly because of role’s burden of each member.
For James March, members must be allowed time where action can be taken before thinking, where intuition plays a preponderant role and where unlearning [from the code] is possible.
