Most Performance Problems Aren't Performance Problems
Investigate before you assume
In the late summer of 2018, I was managing a trailer yard on the 18:00 to 06:00 shift, and we started seeing something I hadn’t encountered before: a string of collisions. Day after day, drivers parking trailers in the yard were clipping the trailers in the spaces next to them. Not dramatically, never any serious damage. But consistently enough that I had to take it seriously.
My first assumption was what most managers’ first assumption would be: something had changed about how my drivers were operating. I started talking to them about being more careful, asking for a spotter when the spacing was tight, looking for spaces with more room before committing to a tight one. I thought I was addressing it.
I wasn’t. The collisions kept happening.
Every shift, I went into the site sync meeting with other department leaders and explained that yes, we’d had another incident. Then another. Then another. By the fourth or fifth incident report in a single week, I had run out of new things to say in that room. Whatever I was doing clearly wasn’t working, and I was out of ideas.
Finally, one of the other department managers asked me a question I hadn’t thought of: what was the lighting like in the yard when the collisions were happening?
That was it. The yard lights were on a timer that had been set earlier in the summer, when the sun went down later. By late summer, the sun was setting earlier—but the lights weren’t coming on until the same time they always had. There was a fifteen to twenty minute window every evening when the yard was genuinely dark and the lights hadn’t kicked on yet. My drivers hadn’t been parking carelessly. They’d been parking in the near-dark.
The facilities team adjusted the light timers. The collisions stopped immediately.
I had been coaching a lighting problem.
The Default Assumption
When someone’s performance drops, the instinct is to look at the person. Are they engaged? Are they putting in the effort? Do they understand the expectations? Is there an attitude issue here?
These are reasonable questions. They’re also, more often than not, the wrong starting point.
W. Edwards Deming, whose influence on manufacturing and operations management has touched almost every warehouse and production floor that takes quality seriously, had a principle that I return to constantly: “Every system is perfectly designed to get the results it gets.”
When I first came across that idea, I read it as being about process improvement. And it is, in part. But the deeper application that I’ve found most useful as a manager is about performance diagnosis. If your system is producing an outcome you don’t want, the first question isn’t what’s wrong with the person producing that outcome. It’s what’s in the system that’s making that outcome possible, or likely, or even inevitable.
This should change everything about how you approach a performance conversation. Instead of starting with the person and working outward to find context, you start with the context and work inward to find the person. You’ll often discover that the person is the last place the problem lives.
Before you decide you have a performance problem, ask yourself whether you might have a system problem, an equipment problem, a training problem, or an environment problem wearing a performance problem’s clothing.
What’s Actually Masquerading as a Performance Problem
These categories aren’t exhaustive, but in my experience they cover the vast majority of situations where a manager is about to have the wrong conversation.
The most literal category is equipment and environment. Yard lighting that doesn’t account for how early the sun sets in autumn. A workstation at the wrong height for someone who isn’t average height. A radio that doesn’t work reliably, forcing someone to make decisions with incomplete information. Background noise at a level that makes it difficult to hear instructions clearly. None of these show up in the data as equipment or environment problems. They show up as collision rates, productivity drops, quality issues, and missed communications. The people experiencing them often don’t connect the dots either or, if they do, aren’t sure whether it’s worth mentioning. And the manager, looking at the numbers, sees a performance pattern and reaches for a coaching conversation.
System and process problems are subtler but just as common. When I wrote about process improvement fundamentals, I mentioned something that bears repeating here: if your system makes the wrong thing easy and the right thing hard, you’ll get the wrong thing, regardless of coaching. A procedure that’s ambiguous at a critical decision point will produce inconsistent results. A handover process that doesn’t transfer key information will produce errors in the shift that inherits the gaps. The diagnostic tell is that the issue follows the process, not the person. When you see the same error type appearing across different associates, or at the same point in a shift regardless of who’s working, you’re almost certainly looking at a system issue. No amount of individual coaching fixes a procedural ambiguity; it just moves the inconsistency around.
Training gaps are particularly common with associates who’ve been in a role for a few months—long past the point where anyone thinks of them as new, but still operating with gaps from an onboarding that covered the standard cases without fully preparing them for the edge cases. They’ve been managing around those gaps and at some point the workaround fails. In last year’s post 7 Early Warning Signs of Performance Issues, I wrote about how performance problems often develop gradually before they become visible. Training gaps follow this pattern too. The associate seemed fine for months because the work was within the range they were prepared for. Then something shifted (volume, complexity, a role change, a new process) and the gap became a liability.
The hardest category to identify from the floor is scheduling and personal factors. A person whose sleep is being disrupted by a shift change they’re struggling to adapt to. Someone whose commute changed and is now arriving with less recovery time than before. A home situation that’s creating sustained background stress. These show up in the performance data as energy drops, attention lapses, and inconsistency that looks, from the outside, like effort. This is also the category where the gap between “this looks like a performance problem” and “this is a person going through something difficult” is most consequential, and where genuine curiosity matters more than any diagnostic framework.
The Diagnostic Shift
In the 5-Why Technique for Warehouse Problem Solving, I wrote about a reframe that changed how I investigate recurring problems: instead of asking “why did this happen?”, ask “why was it possible for this to happen?”
The first question points at the person or the moment. The second question points at the system: the conditions that allowed the error to occur, the gaps that made the outcome possible, the environmental factors that no one thought to examine. It’s a more useful starting point for almost every performance situation, not just operational ones.
Applied to performance, the shift sounds like this: instead of “why isn’t this person hitting rate?”, ask “what would make it difficult for anyone in this position to hit rate?” Instead of “why does this associate keep making this error?”, ask “what in the setup of this role makes this error easy to make?”
Those questions don’t excuse performance issues. They locate them. And locating a problem correctly is the prerequisite for addressing it effectively, something I learned the hard way through a week of embarrassing site sync meetings before a colleague asked me about the lighting in my yard.
Before going into any performance conversation, I now work through four questions:
What changed around this person around the time the issue started? Performance rarely drops randomly. There’s almost always a trigger. A shift change, a process update, a change in equipment, a change in who they’re working alongside, a change at home. If I can identify what changed, I can usually identify what category the problem lives in.
Has this person performed well at this role before, under different conditions? If yes (and for most performance issues in people who aren’t new, the answer is yes) the question is what conditions are different now. If no, the question is whether they were set up for success in the first place.
Have I looked at the environment they’re working in? Literally looked. Have I walked the area, checked the equipment (and the lights), asked whether anything is harder about this position than it should be?
What does the person themselves say is making it hard? People know things about their own experience that don’t appear anywhere in the data. The person who’s struggling usually has some sense of what’s contributing. They may not have connected all the dots, and they may not have felt it was their place to say anything, but they know things. Ask.
What Changed?
The single most diagnostic question in a performance drop is: what changed?
Not “what changed about the person?” What changed in the situation? The shift pattern. The team composition. The physical setup. The process they’re following. The manager they’re reporting to. The pace of the operation around them. The handover they’re receiving.
My systemic approach to this full root cause process is what I covered in The 5-Why Technique for Warehouse Problem Solving. But you don’t need the full technique every time. The minimum viable version is just the question: what changed, and when did it change relative to when the problem started?
When It Is the Person
None of this is an argument against ever concluding that a performance issue is fundamentally about the individual.
Sometimes it is. Genuine effort issues exist. Genuine attitude issues exist. I wrote about The Conversation You Keep Postponing and the shared the structures in the Difficult Conversations Planner exist because they’re needed. A manager who responds to every performance problem by looking for a system to fix is avoiding individual accountability as surely as a manager who responds to every system problem by looking for a person to coach.
The distinction between the two isn’t always clean. What it requires is diagnosis, a genuine effort to understand the situation before deciding what it is, rather than assumption. The assumption almost always lands on the person. The diagnosis sometimes gets there too. But it arrives there by a route that’s more honest and more useful.
The question that keeps this honest is a simple one: if I put a different person in exactly this situation—same station, same shift, same equipment, same handover—would they produce the same issue? If the answer is probably yes, you have a system problem. If the answer is probably no, you have a person problem. The diagnosis flows from that distinction.
Summary
Most performance problems look, from the outside, like something wrong with the person. The numbers drop. The quality slips. The errors repeat. The data points at the associate, and the instinct is to start there.
The better starting point is the system. The equipment. The environment. The training. The process. What changed around this person, and when? What would make it difficult for anyone in this position to perform at the expected standard? What does the person themselves say is making it hard?
Deming was right: every system is perfectly designed to get the results it gets. Before you decide the result you’re getting is a performance problem, ask whether your system might be producing it for reasons that have nothing to do with the person experiencing it.
From Theory to Action
Before your next performance conversation, look at the environment. Physically. Walk the area. Check the equipment and conditions. Ask whether anything about the setup is harder than it should be. You’re looking for the thing that wouldn’t show up in the data: the lighting problem, the equipment problem, the workstation problem. The collisions in my yard didn’t show up as a lighting issue in any report. They showed up as driver errors. Ten minutes of looking at the environment before a coaching conversation is worth more than an hour of coaching a problem you haven’t correctly located.
Ask “what changed?” before you ask “what’s wrong?” Think back to when the performance issue started. What changed in the operation, the team, the process, or the person’s circumstances around that time? If you can’t identify a trigger, that’s important information too. But in most cases, performance drops have a starting point and the starting point is a clue.
Look for the pattern, not just the person. Is this error type appearing with multiple people? At the same point in the process? On a specific shift? If the problem follows the process rather than the person, you’re looking at a system issue. No individual coaching will solve it; something in the setup needs to change. The 5-Why Technique is the right tool for this investigation.
Ask the person what’s making it hard. “I’ve been looking at your numbers and I want to understand what’s getting in the way. What’s making this difficult?” The answer won’t always reveal something environmental. But it will often surface something you didn’t know, something that changes what you do next. People who are struggling usually have some sense of the contributing factors; they may just not have felt it was their place to raise them. My drivers knew the yard was dark, but none of them thought to mention it, and I never thought to ask.
Run the substitution test. Before deciding the problem is the person, ask: if I put a reliable, experienced associate in exactly this situation—same station, same shift, same equipment, same handover—would they produce the same issue? If yes, fix the system. If no, you may have an individual performance situation. Either way, the test tells you where to look.
Separate the diagnosis from the conversation. The Difficult Conversations Planner and the framework in The Conversation You Keep Postponing are the right tools once you’ve determined the problem is fundamentally about the individual. But run the diagnosis first. The conversation you have after correctly locating the problem is more specific, more fair, and more likely to produce something useful — both for the person and for your operation.
Document what you find. Whether the diagnosis leads you to a system fix or a performance conversation, write down what you found and what changed. Your Growth Documentation System is the right place for this. Pattern recognition across multiple situations — seeing that the same station produces problems for multiple people, or that a specific process step keeps generating errors — is only possible if you’ve kept records. The lighting problem in my trailer yard took longer to solve than it should have partly because nobody had documented the fact that the collision rate at that time of evening had been elevated before. The pattern existed. It just wasn’t written down anywhere.