The dirty little secret of the booming agile training and coaching industry is that most Agile implementations ultimately fail, meaning that the companies revert back to their previous working methods. Even ones that experience dramatic early success and drive their entire industry to implement similar approaches typically lose momentum after several years or with a change in leadership. It is telling that, of the first wave of companies to implement Scrum in the mid 1990’s, not one is still using the framework today. In light of this, it is tempting to blame agile principles or the specific execution frameworks themselves for this lack of longevity, but given the extensive track record of short-term success that would be throwing the baby out with the bathwater. I believe the root cause is a narrow focus on the tactical work implementation, and corresponding failure to think of enterprise agility in a holistic strategic context.
One part of what makes strategic agility so difficult to implement and sustain is that, like all good strategies, it requires an internally consistent system of activities that support delivery, reinforce or enhance each other, and require deliberate trade-offs with other strategic approaches. Nikolaj Siggelkow describes these internally-consistent decisions as an “activity system.” Figure 1 below shows a representative activity system for an agile competitor, highlighting the “modules” of activity needed to build and sustain enterprise-level strategic agility. These modules are not the activities themselves, which are unique to the needs of each company…they instead capture the overall goals that the specific activities need to meet to hold the entire agile activity system together. You will notice that each activity module supports one or more stages of the PDCA cycle described in the original agile corporate strategy post.
Figure 1: An example activity system for an agile strategy competitor
The system is anchored by four core activity modules, and reinforced and enhanced by 14 additional modules that knit the activity system into a difficult-to-duplicate whole.
- Iterative and incremental delivery – executing strategy in sequential increments to provide opportunity for feedback and learning
- Distributed decision making – empowering individuals closer to the front lines to make decisions relevant to their operating sphere
- Mindset of continuous improvement – always looking for ways to make things incrementally better
- Short feedback cycles – working to structurally minimize the time between taking action and measuring the results of that action
- Unifying strategic vision as the living skeleton that supports strategic activity
- Customer-centric focus to ensure the entire organization is attuned to market needs
- Modularly scalable systems that make it easier to grow and shrink in response to market change
- Explicit strategy tests to validate or disprove key assumptions baked into the strategic vision
- Cross functional teams that have all commonly needed skills in one closely-coordinating group
- Aligned incentives to avoid unintended adverse consequences of individual decision making
- Autonomous teams that have the freedom to respond in near real time to realities on the ground
- Low change-over cost that makes it easier to switch from one approach to another
- Clear success metrics that are known to everyone and support shared objectives
- Regular internal review to provide opportunities to refine strategy, product design and processes
- Transparency on goals, status, and progress to ensure everyone is making fully-informed decisions
- Engaged customers who can provide thoughtful and actionable feedback on the current state
- Clear prioritization of the steps needed to deliver the strategy so that everyone knows what to do
- Emergent design that develops in parallel with strategic hypotheses, structured experiments and feedback
As mentioned above, an individual company will use their own specific activities to fulfill each activity module, and many of these activities are codified in standard Agile methodologies. For example, the formal definition of Sprint cadence and meetings in Scrum is an internally-consistent set of activities that meet the iterative and incremental delivery, regular internal review, and short feedback loops module requirements above. However, hard-won experience has shown that some form of aligned practice is needed for each of these activity modules to hold the overall system together, which extends beyond any one framework. Importantly, this means that successful strategic agility involves thoughtful adaptation rather than rote implementation of a single branded Agile framework.
When Agile organizational implementations fail, it is typically because of a missing activity in this system, or internal inconsistency between the activities used. For example, a very common implementation problem occurs when organizations want autonomous teams to make distributed decisions, but don’t change previous risk-averse incentive structures that discourage teams from committing to independent decisions. Similarly, an organization that attempts to implement an Agile activity system without a unifying strategic vision of how the company should compete in the market and explicit strategy tests to verify the vision risks falling into the operational effectiveness trap of trying to do the same thing as all its competitors, only faster. This commoditized time-based competition can rapidly compete away any profits from productivity gains. This has proven the case over the past two decades of declining profits at Japanese electronics manufacturers, who use lean and agile delivery methods extensively to reduce cost and speed time to market, but without an accompanying vision of how this agility allows them to be different than their competitors.
Good strategy requires deliberate tradeoffs, and Agile strategy is no exception. The Agile activity system above includes activities that would harm a low cost or differentiated competitor.
- Modularly scalable systems are typically a little over-built relative to monolithic systems to avoid the high cost of redesigning connections between modules. This is just one example of where optimizing for low change-over cost often increases unit costs relative to a high throughput system, which would be disastrous to a low cost competitor.
- Flexible and autonomous cross-functional teams are vital to an organization’s ability to react quickly to updated plans, yet consciously give up the detailed expertise of specialists operating in a classic management hierarchy.
- Embracing emergent design means the organization needs to be comfortable not knowing the ultimate destination when they set out and willing to accept that they may head down some dead ends en route to a successful strategy. This is deeply uncomfortable for organizations used to following a pre-defined plan. (even though the pre-defined plan is almost always wrong)
The good news from a strategy perspective is that difficult-to-implement systems are also hard to duplicate. Because agile implementations that lack a coherent activity system ultimately fail, a competitor must either duplicate the entire activity system or develop their own thoughtful and consistent agile activity system that meets their unique needs and position in the market. This makes a true agile strategy both sustainable and highly defendable from competition.
 Siggelkow, Nicolaj. “Firms as Systems of Interdependent Choices;” Journal of Management Studies; July 2011.
 Stalk, George Junior and Alan M. Webster. “Japan’s Dark Side of Time;” Harvard Business Review. July-August 1993