
Practicing what we hope we will not have to do
Knowing a plan does not demonstrate that we can execute it. Practice allows specific dependencies to be checked, corrected and tested again without promising a permanent guarantee.
A company brings its team together to review what it would do if its order intake system went out of service. In this hypothetical scenario, people know the plan: activate an alternative channel, preserve essential information and tell operations how work will continue. The conversation helps clarify responsibilities and resolve some differences in interpretation. By the end, everyone can describe a coherent response. We still do not know whether they can put it into action.
That distinction does not make the conversation useless. A discussion can reveal contradictions worth resolving before attempting the sequence. In its emergency preparedness materials, FEMA distinguishes discussion-based from operations-based exercises (opens in a new tab): the former allow plans and procedures to be examined; the latter include actions in response to a scenario, such as initiating communications or mobilizing resources. That distinction, situated in emergency management, helps clarify a more modest decision within a company: what we need to understand through discussion and what we can check only by trying it.
What the script does not resolve for us
Preparation for an infrequent situation contains dependencies of this kind. An instruction may remain correct while referring to a tool that has changed; a person may retain their experience while needing an authorization they no longer have. Everyday operations may not test those relationships because they never need to use them. Waiting for an incident leaves unobserved precisely the part of the work that justified having a backup.
Disuse can also affect performance, but should not become an automatic explanation. The meta-analysis by Corey Tatel and Phillip Ackerman published in 2025 (opens in a new tab) examined retention of procedural skills with motor components and found deterioration associated with intervals without use. It identified task type and complexity, intermittent opportunities to perform and instructions as possible moderators. Its estimates do not set an expiration date for professional judgment or explain a team’s readiness on their own. In our exercise, for example, the obstacle was a permission, not a demonstrated loss of skill.
Choosing what to practice requires bringing those questions together without confusing them. What matters is how much harm an inability to respond would cause, which part of the response is rarely used and which conditions have changed since the last check. A system modification may justify reviewing the alternative channel even if the team practiced recently. A capability regularly exercised in comparable conditions may need less additional rehearsal. Frequency should respond to that situation and the applicable requirements, not to a single schedule inferred from an average rate of deterioration.
We also need to decide how much of the problem to represent. Checking the permission in the example does not require simulating a company-wide interruption. If we want to observe how several areas coordinate, an isolated access test falls short. The exercise’s scale needs to relate to the dependency we want to examine and the risk of testing it. Increasing difficulty does not automatically improve observation; it may introduce harms or distractions that do not help us learn.

What we leave changed after the exercise
Practice allows learning during the attempt. It may be necessary to stop, explain a decision or repeat a step with help. That support has value if the purpose is to develop the response. What it does not permit is describing the sequence afterward as though it happened without intervention. When we resume practice, we need to know what was resolved, what support was needed and what we want to observe again. Repeating a sequence until it works is not enough to conclude that we can respond to any variation of it.
The subsequent review can turn that attempt into a more useful experience. Rather than merely asking whether the exercise went well, it allows what was expected to be compared with what happened: where action stopped, what information helped and what made recovery possible. Scott Tannenbaum and Christopher Cerasoli’s meta-analysis of after-action reviews (opens in a new tab), published in 2013, found average performance improvements over control groups across diverse contexts. It supports taking that structured review seriously; it does not allow us to promise a particular improvement in the response to a future real interruption.
In our case, noting that a permission was missing still leaves the problem open. Someone must review which permission should be enabled, within what limits and how to check that the correction allows the intended operation. The team also needs to become familiar with the route it can now use. A new test can confirm that link without thereby certifying the continuity of the entire service. What changed is concrete: a dependency previously assumed could be examined and, once corrected, tested again.
That result may be more valuable than an exercise with no apparent difficulties. Not because finding faults is an end in itself, but because there is now one fewer open question about an important condition. Preparation retains a defined scope: which situation was practiced, who participated, with what resources and what remains unresolved. The absence of incidents during the year does not provide that same information, and a day of rehearsal does not turn it into a permanent guarantee either.
Practice allows us to address some of the conditions we previously took for granted before we need to rely on them. If a critical dependency can be checked through a safe, proportionate exercise, we should not save its first test for the day we actually need it.





