When Christina Gibson and Julian Birkinshaw showed managers could create contextual ambidexterity by creating a culture that expected both innovation and operational performance, it was like injecting oxygen into a very slow-burning fire. Think of contextual ambidexterity as agile for managers.
Robert Duncan birthed the topic of organizational ambidexterity but framed it as a structural separation. Twenty-eight years later, Gibson and Birkinshaw showed that firms had contextual ambidexterity, that contextual ambidexterity was a management capability, and that the path to contextual ambidexterity was the unique product of behaviour and culture in context.
During that period we entered the digital age, and the role of the manager has adapted accordingly.
Ambidexterity is an academic concept - a theory developed by researchers to explain how organizations resolved the core tensions between:
"exploration of new possibilities and the exploitation of old certainties".March, 1991
James March was interested in how people adapted knowledge within organizations so he simulated the process. It is his highly influential paper, Exploration and exploitation in organizational learning, that established the terms exploration for innovation and exploitation for gathering the profits from innovation. The word is now abhorrent, so let's call it operations instead.
Duncan had noticed managers separated innovation and operations and thought it was so they could be ambidextrous - good at both. Organization Design was an emerging topic and Duncan's findings were published in the proceedings of a conference in 1976. Duncan's paper was called 'The ambidextrous organization: Designing dual structures for innovation'.
Structural separation was given, so academics needed to create a theory that explained integration. Mary Benner and Michael Tushman, as well as Charles O'Reilly, held that integration was a function of leadership. Their case studies featured leaders capable of communicating a "unifying vision".
Structural separation and poor integration are common
Innovation and operations activities were traditionally conducted separately, and still are if you have separate R&D departments. Many firms even maintain different budgets (RTB / CTB) and functional hierarchy means separate managers and governance too. Separation allows innovation and operations mindsets and knowledge to prevail in their own context.
The challenge comes when it's time to integrate new knowledge into existing routines. People from each discipline think and behave differently. They have different ways of managing risk and governance that can be incompatible. Collaboration and cooperation skills are often underdeveloped in siloed organizations, and the situation is often made worse by individual incentives. Without contextual ambidexterity, the only theoretical option beyond leadership by top management was dedicated management of handover processes.
Failing to balance innovation and operations leads eventually to failure, just as it did when transistors disrupted the valve industry, and quartz movements the Swiss watch industry. Tushman and O'Reilly's (1996) article told leaders to become Machiavellian revolutionaries. If you can't access that but have access to HBR, you can read their updated (2004) version, The ambidextrous organization.
Get ambidexterity right, and your firm / department / team performance can be outstanding in both areas. Paradox researchers Wendy Smith and Marianne Lewis call this Both/And Thinking in their book, and warn leaders that solving today's tough problems takes a mindset shift in their HBR article.
By 2004, Gibson and Birkinshaw were able to show that separate structures were not the only path to ambidextrous performance.
They established that many business units achieved ambidextrous performance by creating a context where both adaptability and alignment were achieved. The mechanism was for managers to create the appropriate culture and clarify the priorities, in other words, to expect both innovation and operation to take place and support everyone in reaching those goals. Obviously, contextual ambidexterity has a lot in common with agile. Adaptability and alignment are not quite the same as explore-exploit or innovation-operation, but they are opposing forces. This is how they define them:
"Adaptability: the ability to move quickly toward new opportunities, to adjust to volatile markets and to avoid complacency. Alignment: a clear sense of how value is being created in the short term and how activities should be coordinated and streamlined to deliver that value."Gibson and Birkinshaw, 2004
DevOps is an ambidextrous solution pattern
In 20th-century IT, Development was all about change, and Operations was about preserving reliability. Everyone believed it was either one or the other.
With DevOps, we understood that by making both change and reliability the priority, performance actually improved. In one investment bank I helped, we both doubled the rate of software releases and halved the number of incidents every year for six years. Nicole Forsgren, Jez Humble, and Gene Kim provided the evidence Tech leaders need to make an informed decision on this in their book, Accelerate.
By 2021, the annual State of DevOps survey (pka DORA) reported that high-performing DevOps organizations deployed almost 1000 changes to production for each change made by low performers. Their finding that it took low performers 6,570 times longer to deploy their code changes to production was so outstanding they added the text:
"Yes, you read correctly. This is not an editorial error."State of DevOps Survey (2021) Google
Don't forget, DevOps is an IT case and academic research is interested in strategy and theory that applies to whole organizations. This is why I'm so interested in contextual ambidexterity as a model for agile transformation.