Reversibility is one side to the implementation of a solution. Another aspect to consider is the impact of the worst case scenario. Your course of action would be different when a life-threatening situation could occur compared to a situation where you would spend two hours of your life without results. Combining the two factors yields the speed-reversibility matrix. It dictates the speed with which to move through the decision-making process taking reversibility and impact into account.
A decision can either be reversible or irreversible, and the decision can have a high impact of the worst case scenario (also called ‘high risk’) or a low impact (or ‘low risk’). All possible combinations of these two factors are decisions that are reversible-high impact, reversible-low impact, irreversible-high impact and irreversible-low impact.
Figure 1. The speed-reversibility matrix has decisions that are reversible-high impact, reversible-low impact, irreversible-high impact and irreversible-low impact.
Reversible-high impact decisions push you towards the multiple small experiments mode. As each decision can be made undone, you can repeat the decision with different parameters giving different results. After experimenting with several scenarios, you can make a decision on which scenario gives the optimal result and go with that solution.
Reversible-low impact decisions are entertaining as they enable you to move fast. Why bother gathering much data to analyse the solution, as you can just get started with a solution that cannot have disastrous results. In case you made an error, you can simply revert the operation and try again. The real difficulty with these problems is in recognising them correctly and not labelling them as one of the other decision-making patterns. You will lose speed by incorrectly judging the problem as irreversible and acting accordingly.
Irreversible-high impact decisions ask for slow and detailed consideration. The option you take is your definitive choice, and a wrong move can lead to disastrous results. It is paramount to collect and analyse as much data on the problem as you can find. Maybe you can devise reversible-high impact scenarios that closely resemble this irreversible problem and do several small experiments to give you a feel of the right course of action. You might come to the conclusion that you’re not able (yet) to make the decision now. Simply wait, gather more data and make the decision later.
Irreversible-low impact decisions allow for errors to happen, but cannot be made undone. Here, it is best to gradually roll out your solution and monitor what happens. When the coast seems clear, roll out more steps. When an error occurs, analyse what went wrong and if this affects the next step of the rollout. Adjust if necessary and continue the rollout process.