Agile Scenario Planning

'The future is not fixed.'

'There are many possible futures before us.'

'The future does not need to be like the past.'

All truisms, but sometimes hard to live. We all get trapped into linear narratives, believe that past trends will determine the future, and find it hard to hold in our imaginations several different possible worlds. People who can imagine multiple possible futures have an advantage when it comes to sensing and responding to change and shaping better outcomes.

One way to hold different futures in your mind at the same time is Scenario Planning. This is a mature business process that has been used many times in many contexts. At Ibbaka we sometimes use it to help our clients develop pricing strategies that will be robust in different economic and competitive contexts. See The Jobs of Pricing Scenarios.

One of the early case studies in scenario planning is also one of the most compelling, the Mont Fleur scenarios. This work was led by Adam Kahane back in 1991 and was intended to help leaders in South Africa move beyond apartheid. You can read more about this important example of the Reos Partners website.

The classic approach to scenario planning

The classic approach to scenario planning generates four scenarios based on combining two critical uncertainties to generate four scenarios. There are more complex approaches with more critical uncertainties and more scenarios but these are not as common and in practice are more difficult to use. Each of the scenarios is a rich story supported by data and images. Ideally early indicators are provided that help one tell which scenario is emerging.

There are several possible uses for these scenarios.

• Find a desirable future and guide people to that future (like the Mont Fleur scenarios)
• Make investments that are viable across multiple future scenarios (this is often done in urban planning)
• Be prepared to recognize and respond to a scenario quickly (as in the classic example of how Shell was able to respond quickly to the 1973 oil shock)

There are many online resources and books that can give you more details. See for example the Scenario Thinking Wiki. A good introduction to the broad literature on scenario planning is Scenario Planning: The Link Between Future and Strategy by Mats Lindgren and Hans Bandhold.

A simple process for the classic approach to scenario planning is summarized below.

Set context

1.What is the purpose of the scenarios?

2. Who will be involved in developing them?

3. Who will use them?

4. What is the time frame (this is usually three to twenty years)

Under the trends and critical uncertainties

5. What are the trends over the time period we can be relatively certain of?

6. What are the critical uncertainties that could go either way and would have a significant impact on the future state?

Creating the scenarios

7. Map out dependencies between critical uncertainties (this is often skipped but it is important to understand these dependencies)

8. Work through each combination of two critical uncertainties and see which generate the most 'meaningful' scenarios (meaning will depend on purpose)

9. Develop the scenarios by creating stories and images about each of the different possible futures; identify leading indicators that will suggest how each of the critical uncertainties is resolving (what scenario is most likely to come true)

Tracking the critical uncertainties and taking action

10. Prepare action plans for each of the scenarios

11. Track the leading indicators

12. Review the scenarios regularly with all stakeholders

Limits to the classic approach to scenario planning

Classic scenario planning has proven compelling in many contexts, but it does have its limitations.

The biggest obstacle is the investment required to develop and maintain a set of relevant scenarios. Depending on how many people are involved and how deep the research, this can take anywhere from three months to a year or more and require a dedicated team supported by consultants. Many organizations choose not to make this investment and when they do fail to maintain the scenarios.

Another challenge can be the simplification forced by having only two (or at the most three) critical uncertainties. In today's dynamic world there are often more critical uncertainties at play.

A third issue is the possibility of missing a critical uncertainty. Very few organizations had the social and economic dislocations of a global pandemic or a Russian invasion of Ukraine on our radars (though perhaps we should have). We need a way to rapidly surface and factor new critical uncertainties as they are emerging.

Classic scenario planning is a powerful framework, but we need an alternative. One that is lighter and faster, one that allows for a greater variety of uncertainty, and that can sense and respond to a changing environment more quickly.

Agile Scenario Planning

When the Covid pandemic broke relatively few organizations were prepared or had a scenario ready for a global pandemic. This is a bit surprising as we knew, or should have known, that a pandemic was inevitable at some point and we had the early warning of SARs in 2002-2003. But none of the scenario plans I have been able to review, across about 20 organizations, had the emergence of a pandemic as a critical uncertainty. There were too many other things that seemed more directly relevant to worry about (on the other hand, many scenarios have good coverage of climate change, see for example Shell's Energy Transformation Scenarios).

Agile Scenario Planning shifts the focus from the scenarios to the critical uncertainties. It keeps itself open to the emergence of new critical uncertainties and how they could combine to generate unexpected effects.

Rather than constructing four (or six or nine) formal scenarios, a list of critical uncertainties is developed and kept updated. A process is needed to keep the list current and to detect how each critical uncertainty is resolving. Depending on the cadence one might need to check the trends hourly, daily, weekly or monthly. If you can wait and check on an annual basis then the classic approach to scenario planning is likely more relevant. The process also needs to be able to generate new critical uncertainties as they emerge.

Combinations are still important in agile scenario planning. The power of the scenario planning approach comes from combining critical uncertainties to generate scenarios and the often surprising interactions that arise. Emergent phenomena, and even self organization, are generally the result of combinations.

The challenge of agile scenario planning is managing the combinatorial explosion. If one is tracking 5 critical uncertainties there are 10 combinations to consider (assuming one is limiting oneself to binary scenarios based on only two critical uncertainties, there are more nuanced approaches). If ten critical uncertainties are being tracked the number of combinations grows to 45, 20 critical uncertainties generates 190 combinations. This rapidly becomes unmanageable.

(The formula for calculating this is N(N-1)/2.)

How can we thin down the number of combinations to consider in agile scenario planning?

The first filter is to understand causal dependencies between critical uncertainties. This can be part of the classical process but in fact few people bother, it is essential in agile scenario planning. Related to this is the elimination of combinations that include mutually incompatible scenarios. To take a silly example, 'Sea Level Rise' and 'Global Temperature Decline' do not need to be considered together.

One tool for mapping the causal relationships between uncertainties is the design structure matrix (DSM). Originally developed in the 1960s to manage systems of equations, DSMs became popular in the 1990s to organize complex products and improve modularization. DSMweb.org has a large number of examples and tools. I have written about some uses of DSM in skill management here.

To apply DSM to critical uncertainties you just list each uncertainty on the X and Y axis and note how the might influence each other. If you don't know then you need to dig in and develop your understanding. A simple example is shown below.

There are other approaches of course. One could also use systems diagrams or Judea Peal style causal diagrams to explore the relationship between critical uncertainties. But these are advanced techniques, not relevant in all situations.

(Figure 10.2 from The Book of Why, Judea Pearl (2018))

Agile scenario planning needs early warning systems. Agile scenario planning is meant to help improve response to fast changing situations. It is critical that one have early warning systems in place to let you know how critical uncertainties are resolving. These can take the form of data alerts, monitoring social chatter, conversations with expert networks and so on. There is a growing array of ways to sense change, but these need a structure to make sure that the signal is not lost in the noise. Agile scenario planning can provide that structure.

Develop your responses to different scenarios before they appear. Agile scenario planning is most relevant when time is of the essence. In these situations, it often helps to have a planned response to key combinations of critical uncertainties. A portfolio of responses to consider for different scenarios as they emerge is an important part of this approach. By hashing out the possible responses before hand one can act quickly.

A Simple Process for Agile Scenario Planning

The process for agile scenario planning is different from classic scenario planning. Whereas the latter is focussed on creating durable scenarios and telling stories that will prepare people for change, agile scenario planning is much more about exploration and preparation of rapid response.

1. Define the purpose (be prepared, explore potential, generate change)
2. Set cadence (daily, weekly, monthly)
3. Identify initial set of critical uncertainties (start with three to seven)
4. Map dependencies between critical uncertainties (to deepen your understanding and manage the combinatorial explosion)
5. Generate combinations of critical uncertainties (based on the dependencies identified in step 4, but include some 'random' combinations that could surprise you; note that you do not need to develop these or go deep, you often want to regenerate combinations and have fresh conversations)
6. Prime your early warning system (what data, sensing, conversations will help you know early how each critical uncertainty is resolving)
7. Prepare your initial responses (this should be done for both critical uncertainties and for scenarios generated by combining critical uncertainties)
8. Have cadence meetings - at each meeting make sure to

• Look at early warning status for all critical uncertainties
• Consider generating (or swapping out) new critical uncertainties
• Review 3-5 combinations from the set of possible combinations
• Review how to respond to these 12 to 20 different scenarios
• Identify scenarios where response would be inadequate
• Review cadence (does it need to be accelerated, can it be slowed down)

That may seem like a lot of work, but ask yourself, what is the cost of being blindsided. The shocks of the past few year have led us to appreciate that resilience and adaptation are as important as efficiency (see Managing the tension between resilience, adaptation and efficiency). Agile scenario planning can increase resilience and accelerate adaptation. They are a powerful way to learn how to respond to change.