5 reasons to rethink the 5 whys
Humans are at the core of this exercise, and we have to put humans first
By Madeleine Reese
There are few topics as hotly debated as retrospectives. Even the language we use to describe them shifts to accommodate the varied ways in which organizations—and professional communities—grapple with this difficult work.
From my perspective, the foundation of every organization, every process, is human—and humans are beautifully complicated, surprising, and emotional. This is good! Humans working together have infinite, compounding potential to learn and improve. This also means that we should take as much care with the human experience of systems, of crises, of decision-making as we do with their outcomes. That means prioritizing psychological safety, especially during our post-incident retrospectives and reviews.
Yet the widespread technique for post-incident analysis—the 5 Whys—often steamrolls the human experience of incidents and of systems. Its deceptively simple template—repeating the question “Why” until a cause is found—doesn’t do justice to the complexity of the task. Here’s why I want us to rethink the 5 Whys:
1. The 5 Whys don’t go deep enough.
Teams who use the 5 Whys often stop their analysis at the proximate cause (or causes) instead of diagnosing down to underlying factors: the technique doesn’t dig deep enough. Where the proximate cause can often be expressed in a simple phrase like, “The documentation was not updated,” underlying factors are behaviors, points of view, or habits—for instance, “There is no clear owner for updating the documentation.” Identifying the varied underlying factors for any incident is an opportunity for innovation, learning, and looping as we consistently evolve our systems.
The openness of retros can be both beautiful and painful: beautiful in what we can learn about ourselves, but painful to honestly face our own weaknesses with vulnerability. What’s scary—and emotional—about them is the underlying issues are always a matter of people, or processes made up of people. Humans are at the core of this exercise, and we have to put humans first. When you only ask why five times, you rarely get those meaty—and profoundly human—answers.
2. The 5 Whys can be arbitrary.
While the 5 Whys can feel like a robust technique for post-incident analysis, its central premise is arbitrary. Why five questions? You may need only three, or you may need far more than that. Moreover, you can get different answers each time, discrepancies that this technique offers no guidance for resolving. Ultimately, because the 5 Whys does not have a goal-oriented design, systematic it will always risk being arbitrary—or outright wrong.
3. The 5 Whys limits diverse thinking and perspective.
Because the 5 Whys is so focused on the question “Why?”, it doesn’t lean into the vulnerabilities at play during post-incident review. While I strongly agree that post-incident analysis should not be about blame or credit, the idea of a “blameless” review often masks the varied underlying whys, which are consistently about people.
After all, an organization is a complex system made up of humans—humans making decisions, taking actions, behaving in messy, beautifully interesting ways. It’s humans who carry out the work, or design what’s automated and how. When you dig down to underlying factors, they’re usually a matter of individuals not having access to the adequate training, resources, or systems required to meet a given goal—or even a mismatch with the person assigned to the task at hand. (We all have our strengths and weaknesses!)
The 5 Whys is a protective shield that ends up hurting us as individuals and as organizations: we need to have thoughtful, vulnerable, emotional discussions about why incidents happen. Rather than rely on a limited framework for analysis, organizations should invest in creating a safe space for its teams to explore with open minds what happened and what to do about it. Whatever system you do use for post-incident analysis needs to take humans—irrational, ego-fueled, biased, blinkered, as well as brilliant—into account.
4. The 5 Whys is not an adequate replacement for post-incident analysis training or preparation.
Good post-incident analysis is both difficult and necessary, and it requires advance preparation and training—especially for the person running the review. In many organizations, the 5 Whys isn’t a tool in a well-prepared incident commander’s toolkit; it’s a crutch for someone who hasn’t been offered adequate training.
Post-incident analysis shouldn’t be reduced to a checklist. Hospitals famously have checklists for medical teams going into surgery—they also have literal postmortems. But medical teams reviewing the case of a patient who has died don’t rely solely on a checklist when it comes to analysis. They review what should have happened, what actually happened, the differences between the two, and the behaviors behind those differences.
We, too, need frameworks that take that complexity into account: processes that help expose peoples’ mental maps, biases, and what they don’t know; processes that collect these stories and surface patterns and outliers; processes that create real understanding of how we can evolve and grow together as a team or an organization.
5. The 5 Whys is a linear, narrow process for a nonlinear, complex task.
The 5 Whys is a linear framework that attempts—and fails—to describe a complex, non-linear incident. Usually there are multiple different underlying factors (overlapping, interdependent inputs and behaviors at play) rather than just one. The process started by the 5 Whys ends with a single answer. It doesn’t make space for a full, accurate representation of reality—and it likely won’t achieve the goal of determining what happened and why.