Six Leadership Practices for the Problem Your Corrective Action Didn’t Fix

You updated the procedure, you ran the retraining, you documented the corrective action and closed it out. Then the same deviation shows up again.
If that sounds familiar, I have a question for you: what if the problem isn't the procedure?
I gave the Day 1 keynote at ChainReaction 2026 in Orlando on October 8. "See the Signal. Lead the Change" was written for quality and food safety teams dealing with problems that technical fixes keep missing. The conference theme was Lead with Insight. Scale with Confidence. My keynote focused on what happens in between. Once the data shows you a problem, what does it take to act on it?

TL;DR

  • When the same problem keeps coming back, entertain the possibility that the problem isn't the procedure, training, or lack of oversight. 
  • Technical problems have known solutions. Adaptive problems need shifts in behavior, mindset, relationships, priorities or values. Recurring problems often have both technical and adaptive elements.  
  • Digital platforms help you see the problem. Understanding it and changing the conditions behind it takes leadership.
  • My six practices: diagnose before prescribing, learn from resistance, keep people engaged when pressure rises, put responsibility close to the work, align goals, rewards and expectations, and protect the change when pressure returns.
  • You don't need all six on Monday. Pick one and start there.

When the same problem keeps coming back

I polled the room: when a recurring problem shows up again, what is your first instinct? The options were to fix the process or retrain the team, add more controls or oversight, look for what may be driving the pattern, or bring others in to understand what needs to change. The first two are technical responses. The last two open up the diagnosis. They help you find out what kind of work the problem calls for. There was no trick answer. I was asking about instinct.
Reaching for the procedure or the training first makes sense. Plants run on systems, controls and repeatable processes, and a technical fix is one you can document and audit. But one reframe anchored the rest of my session:
When the same problem keeps coming back, we have to at least entertain the possibility that the problem isn't the procedure, training, or lack of oversight.

Technical problems and adaptive problems

I drew on the work of Ronald Heifetz and his colleagues, who distinguish between technical problems and adaptive challenges. 
A technical problem has a known solution. A sensor breaks and you replace it. A step is missing and you add it to the procedure. The fix draws on expertise and skills you already have.
An adaptive problem has no easy fix. Behavior, mindset, relationships, priorities or values have to shift. Adaptive leadership is about mobilizing people to adapt when there's no clear solution.
Technical work still counts. Management and compliance matter enormously, and sometimes management alone isn't enough. Recurring problems often have both technical and adaptive elements.
Take a sensor alarm that keeps going off. If the sensor is faulty, replacing it is a technical fix. But if the sensor works and people hesitate to speak up because they’re worried about what happens when they stop the line, a new sensor changes nothing. The adaptive work is changing the conditions that make it hard to speak up.

See it, understand it, change it

I built the talk on three steps. Digital platforms help you see the problem. Understanding what's driving it is the part technology can't do for you. Then you change the conditions that keep producing it. My six practices work mostly in those last two steps.
Data gives you visibility. Understanding takes more work.
I assume you already do some of these. Think of them as six moves to reach for when the data says something isn't working and the obvious response isn't creating lasting change.

Practice 1: Diagnose before prescribing

Most leaders are rewarded for solving problems, so when you see one, you act. The danger is prescribing a solution before you've diagnosed the problem.
I borrowed Heifetz's balcony and dance floor metaphor. On the dance floor, you’re surrounded by activity. In a plant, that might mean responding to an alarm, getting the line running again, or handing off to the next shift. You can see what’s happening right in front of you and feel the pressure to make a decision. It’s harder to see whether the same thing keeps happening across shifts or under the same conditions.
From the balcony, you can step back and see those patterns. Your digital platform can help you spot them, and the people doing the work can tell you what was happening in the moment and what options they felt they had. Diagnosis takes both views.
A signal tells you where to investigate. It doesn't tell you the cause.
What to do:
  • Zoom out from the latest incident.
  • Look past the technical failure to the conditions around it.
  • Test your diagnosis with the people closest to the work.
A solid root cause analysis is the technical tool for this. The leadership part is noticing when yours stops too early.

Practice 2: Learn from resistance

When people push back on a corrective action, the instinct is to treat the resistance as something to overcome. Sometimes we label the people: difficult, negative, not bought in. I asked leaders to get curious instead.
Resistance is data.
The question is what it’s data about. Resistance can tell you what people fear losing. It can also surface a practical problem with the plan that you’ve missed. Ask what’s behind the pushback before you decide how to respond.
What to do:
  • Ask what people are trying to protect.
  • Listen before deciding resistance means ignorance or defiance.
  • Use what you learn to improve the plan without abandoning the purpose of the change.
You can listen to resistance and still lead with conviction.
Your CAPA record probably captures the action you took. Record the concerns people raised, what you learned about the conditions on the floor, and how that informed the corrective action. 

Practice 3: Keep people engaged when pressure rises

Change is easy when everything else is going well. The real test comes when production is behind, staffing is short, costs are rising and the customer is waiting, or three of those hit at once.
My metaphor for this is a pressure cooker. With too little pressure, the issue drifts: everybody agrees it matters and nothing really changes. With too much, people protect themselves, disengage or create shortcuts. Your job as a leader is to keep the pressure in a productive range where people can stay engaged in the work.
I polled the room again: when operational pressure rises, which pattern do you see most often? Standards hold and focus sharpens, workarounds start to appear, leaders adjust priorities or resources, or it varies by shift, team or supervisor. Ask yourself what happens to your quality standards when the pressure is on, and whether the answer changes with the shift or the supervisor.
What to do:
  • Name the competing priorities instead of pretending they aren't there.
  • Adjust the pressure so people can stay engaged and own the response.
  • Don't lower the standard just because the tension is uncomfortable.

Practice 4: Put responsibility close to the work

One of the easiest traps for senior leaders is owning a problem so completely that the people who live with it become recipients of the solution instead of authors of it.
My sweet spot is close enough to influence and distant enough to lead. Giving the work back doesn't mean leadership walks away. The people with the knowledge and proximity help shape the response.
What to do:
  • Bring the people doing the work into the diagnosis early, before you've designed the solution.
  • Give them meaningful authority to act, and back them when they use it.
  • Stay accountable for the conditions, resources and priorities they need to succeed.
You can give the work back without giving away accountability.
If you're a quality manager who doesn't control staffing or scheduling, start with the first item. Build the case for why frontline input belongs in the diagnosis before the solution gets designed.

Practice 5: Align goals, rewards and expectations

You can tell people all day what matters. They pay close attention to what your organization actually measures, rewards, promotes and penalizes.
Culture shows up in what the system makes sensible to do.
The graphic I showed highlighted six patterns that can undermine food safety. Speed takes priority over safety. People are penalized when there’s an opportunity to learn. Speaking up feels risky. Promotions reward tight control over developing people. Findings are treated as failures. Inspections and audits become the target.
If you say food safety comes first but reward speed above everything else, you've sent a message. If findings count as failure, people learn not to find things. If speaking up carries a price, silence becomes rational. These are system problems, and people follow what the system rewards.
Closing a corrective action doesn't mean you corrected the problem.
Track whether the problem comes back, as well as whether the prescribed activity got done.
What to do:
  • Compare what you say matters with what you measure and reward.
  • Ask what behavior the current system makes rational.
  • Align goals, measures and rewards with the behavior you say you want.

Practice 6: Protect the change when pressure returns

Even a good change effort meets a test. A production deadline, a major customer request, or pressure from a senior leader can make it tempting to say, “Just this once.”
Sometimes the pressure is workload. Sometimes it's status, hierarchy or the discomfort of challenging someone with authority.
The exception you tolerate under pressure has a funny way of becoming the new standard.
What to do:
  • Decide ahead of time what can't be traded away.
  • Back the people willing to raise the concern or hold the line.
  • Watch what happens under stress, and watch it when conditions are easy too.

Pick one

I didn't want the ChainReaction audience to leave Orlando with six things to implement Monday morning. I asked each person to pick one: which of the six practices would make the biggest difference if you became more intentional about it? Whichever one you chose, start there.
My closing idea was simple: see it, understand it, change it. Your digital platform helps you spot problems sooner by bringing together information that used to sit in separate places. Leadership has to understand the pressures, incentives, behaviors and relationships underneath the signal, then change the conditions producing the number.
A green dashboard is most valuable when it reflects a healthier system behind the numbers. Seeing the signal is the beginning. Leading the change is the work.
So, which practice are you starting with?

Frequently asked questions

It may not be the procedure. Closing a corrective action doesn't mean you corrected the problem, so track whether the problem returns. Then look at the conditions around the technical failure: the pressure people are under, what the system rewards, and whether speaking up carries a price.
A technical problem has a known solution that draws on expertise and skills you already have, like replacing a faulty sensor. An adaptive problem has no easy fix, because behavior, mindset, relationships, priorities or values have to shift. Recurring problems often have both technical and adaptive elements. 
No. Management and compliance matter enormously, and sometimes management alone isn't enough. Your technical controls stay. The adaptive work goes after the conditions that keep producing the failure.
Seeing the signal means using your data, and your digital platform, to spot the problem sooner. A signal tells you where to investigate, and it doesn't tell you the cause. Leading the change means understanding the pressures, incentives, behaviors and relationships underneath it, then changing the conditions that produce the number.
Diagnose before prescribing. Learn from resistance. Keep people engaged when pressure rises. Put responsibility close to the work. Align goals, rewards and expectations. Protect the change when pressure returns.
Ask what’s behind the pushback. People may be worried about losing something, or they may see a practical problem with the plan. Use what you learn to improve the plan without abandoning its purpose.
If the answer changes by shift or supervisor, that's revealing. It may mean your organization relies on individual leadership behavior and not on conditions built into the system. Start with Practice 3: name the competing priorities, adjust the pressure so people can stay engaged, and don't lower the standard just because the tension is uncomfortable.
Yes. For Practice 4, start by building the case for why frontline input belongs in the diagnosis, before the solution gets designed.
Pick one. Ask which of the six would make the biggest difference if you became more intentional about it, and start there.
ChainReaction 2027 is October 18 to 20, 2027, in San Diego. Details at chainreactionsc.com.

Alan Tart

Principal Consultant at Tart Strategic Consulting Group

With nearly 30 years at the U.S. Food and Drug Administration (FDA), including senior leadership roles, Alan brings firsthand experience in regulatory decision-making, organizational leadership, and complex systems change. He works with leaders to address difficult challenges, strengthen their organizations, and move meaningful change forward.