Interventions in a complex world
In a complex world, there is no reliable relationship between intervention and effect. Rather, an intervention is an experiment: the effect may or may not be the expected one.
A prerequisite for dealing with this productively are
- System thinking: interrelationships are not linear, but follow feedback cycles
- Interventions are experiments
In a changing world, it is necessary to keep up with change; if possible, to be faster and drive change. Therefore: experiments.
Why experiments
We make our experiments "safe to fail" so we are not afraid to conduct more of them. A journalist once asked Jeff Bezos, CEO of Amazon, about the Amazon FirePhone, a total flop. Without missing a beat, Mr. Bezos replied, "We are working on much bigger failures right now." Mr. Bezos runs Amazon like a venture capital fund, which only needs to succeed big on a few bets in its portfolio to easily pay for all of its failures. Mr. Bezos is proud of Amazon being a safe place to fail. When we get stuck or aren't learning enough, we take it as a sign that we need to run more experiments. Speed is key with this principle. We don't wait long periods of time before learning that something isn't working. We fail fast and quickly move on to new experiments. Experimenting & learning rapidly helps us achieve continuous improvement.
What are good experiments
Experiments give quick feedback, cost nothing or little, no one has to agree, and it’s easy.
Before starting an experiment, it is helpful to answer these 9 questions:
- What factors determine the current situation?
- Which factors can be controlled or influenced?
- Why should this experiment happen, what question/s can be answered with this experiment?
- What can be observed about the current situation?
- How can you tell that the experiment is moving the situation in the desired direction or producing pleasing side effects?
- How is observable when the experiment moves in an undesirable direction with undesirable effects?
- What is the natural time period of this experiment? When will the results be in?
- If the experiment makes the situation worse, how can the situation be resolved?
- If the situation gets better – how could the experiment be expanded?
The key to good experimentation is finding something you can try now - without budget, permission, or higher providence. Your goal is to test tiny changes to the system to get quick feedback, not to find the all-encompassing solution.
The Paltchinsky principles for experiments
The Russian engineer Pyotr Ioakimovich Palchinsky had a leading role in the introduction of Taylorist principles into Russian industry after the February Revolution of 1917.
He condenses 3 simple rules by which experiments should be designed:
- Variation: trying out new ideas and things
- Survivability: Experiments should be on a scale where errors are survivable.
(This is also why stories should be small. Also, to be able to complete them within one iteration – but mostly because they can also be wrong). - Selection: seek feedback and learn from mistakes.
Decision-making process to maximize options
or: the last responsible moment
For each decision, identify the options available.
Identify the latest possible point in time at which a decision can be made, i.e. the conditions that must be met in order to make a commitment.
Decision time = deadline – option implementation time.
The first decision is made before the first option expires.
Until that moment, you continue to search for new options and refine or expand existing options.
Identify option(s) for each state and be clear in advance about which option should be exercised based on a particular condition.
Try to postpone the decision time.
Most of the time it costs nothing or little.
To do this, we must be able to implement the option as quickly as possible. In the buffer time, try to speed up the process.
Understand that cost optimization is not the same as revenue optimization or risk reduction. Sometimes it pays to invest in more than one option, even if it may cost a little more. Options have value.
Wait with the decisions… and wait… and wait… until the conditions are met.
If you must act, … do it as soon as possible. You can be sure because you know that you will have made the best possible decision.