Leadership is making something happen that would not have happened on its own, or achieving an outcome that is in doubt. Compliance Leadership is the method for getting your people to follow your system, your policies and procedures, and keep following them.

TL;DR: Compliance leadership is the method for getting your people to follow your system, your policies and procedures, and keep following them. It isn’t what most people mean by accountability. It’s how you get your system followed without confrontation, without counting on a bonus plan, and without continual “telling.”
Key Takeaways
  • Leadership is making something happen that would not have happened on its own, or achieving an outcome that is in doubt.
  • Compliance Leadership is the method for getting your people to follow your system, your policies and procedures, and keep following them.
  • The system is measured on results. The people are measured on compliance to the system.
  • Managers design and implement the system, so the results belong to them.
  • Making sure the system is followed means checking. You ask questions, and you go look.
  • Drift is a checking failure.
  • Don’t ask for buy-in. Ask if they’re willing to try.
  • Policies, procedures, and measures are how you drive behavior. People follow the measures on their own, so make sure yours pull them the right way.
  • Pay people for their best work every day, instead of counting on a bonus plan to get the system followed.
  • Start with one standing policy, procedure, or measure. Ask about it every day for a month, then weekly, and don’t ever stop asking.

Let me start with leadership. My definition of leadership is making something happen that would not have happened on its own, or achieving an outcome that is in doubt.

So what does that have to do with getting people to follow a system? Compliance doesn’t happen on its own. You can build a good system, explain it, and train people on it, and most of the time they still won’t follow it. Or they’ll follow it for a while, and then they’ll stop. Getting them to follow it, and keep following it, can be a challenge. Making that happen takes leadership.

Compliance Leadership is the method for getting your people to follow your system, your policies and procedures, and keep following them. It isn’t a new type of leadership, and it isn’t what most people mean by accountability. It’s the how.

The how starts with who is responsible for what. The system is measured on results. The people are measured on compliance to the system. We don’t measure the people on the floor on results, because it is wrong to evaluate people on an outcome they don’t know how to achieve and can’t directly affect. Managers design and implement the system, so the results belong to the managers.

The managers then keep checking that the system is followed, because as soon as you stop asking, people think it’s not important anymore. Managers checking is also how we make sure everyone understands, and how both the system and the people keep improving. What a manager finds when checking is usually one of two things. Someone didn’t understand what the system was asking, so you train them right then. Or the system needs fixing, so you fix it.

So isn’t that just accountability? I don’t think it is. When most people say accountability, they mean who answers for it when the result misses. Compliance Leadership is about how you get the system followed, so the result happens. Here is how the two differ:

  • Compliance Leadership is bigger than accountability. Holding people accountable is one piece of Compliance Leadership, the checking. Compliance Leadership also covers getting the system right, making the right way the easy way, starting it, improving it, and growing your people.
  • Compliance Leadership comes before the miss. Accountability usually shows up after a result is missed. With Compliance Leadership, you’re checking while the job can still be saved.
  • With Compliance Leadership, responsibility goes up. Accountability usually pushes a miss down onto the person. With Compliance Leadership, a miss in the results belongs to the people above them.

So this is not just another article on leadership or accountability. By the end of it, I want you to know how to get your people to follow whatever you’re trying to make stick, whether it’s a scheduling system, a procedure, a quality initiative, or a project. I’ll also cover what others have said about leadership and accountability, and how it all ties together.

My views are shaped by being a Theory of Constraints expert, and by learning from and working with Dr Eli Goldratt, who created it. Goldratt didn’t really talk about leadership or accountability by name. He talked about managing people from a high level: measures, empowerment, respect, giving people the why, and how people judge change. His approach was to use logical thinking so that people communicate clearly and understand each other. He said that everything known about managing people “all boils down to respect,” and that the conclusions of The Seven Habits were correct, but something was missing: the how.[1] Most of the work, he said, had already been done, so he focused on what he felt was missing: how to criticize an idea without hurting the relationship, mismatches between responsibility and authority, how to give instructions, and how to build a team.[2] Wherever he had a position, I’ll give it to you in this article, because I agree with most of it. But none of it tells you what to do when your people have the authority and the know-how, you’ve explained the why, and they still don’t follow the system. There needs to be more meat on those bones, so I developed the Compliance Leadership Method.

Let’s start with the problem.

Your People Won’t Follow the System

You built a system. Maybe it’s a scheduling system. Maybe it’s a new procedure, a quality or safety initiative, or a project that needs to get to the finish line. You explained it, and it is not being followed. Instead, your people are following something else. Here is what I hear from owners of custom job shops:

  • A bonus plan is getting units out the door, just not the units anybody wants right now, and the late list grows.
  • An operator follows the procedure to the letter, can see the job is going to miss, and says nothing.
  • Workarounds are everywhere, and you can’t tell whether a workaround is defiance or a sign the system needs fixing.

So you explain the system again and again, and you are worn out from explaining.

Explaining it one more time won’t close the gap, because the gap was never about knowledge. A shop owner told me he gives all his people a quiz every quarter to make sure they know how the system works. Everybody always passes, and he still battles getting them to follow the rules every day.

Knowing is not doing.

Other times, the system worked, and then it stopped. The Velocity Board got run the way it was designed, the rules got followed, and the results showed up. Then the champion left, or got promoted, or the whole shop got slammed for a month. A year later the board is still on the wall, nobody runs it the way it was designed, and the old chaos is back. In my experience, the Velocity Scheduling System (VSS) sticks for most shops. When it doesn’t, a personnel change usually came first.

Either way, owners end up with the same wish. Shop owners have told me, “I wish they would just comply,” which means, I don’t want to have to stay on top of this, babysit it, or police it. Nobody wants to spend their time staying on top of people.

My clients struggle with this. When a client tells me their people won’t follow the system, I think, yes, your people are resistant, but you are not managing the compliance. That thought is what started me down this route. So let’s look at what owners usually try, and why those attempts aren’t enough.

What Owners Try, and Why It Isn’t Enough

Problems like these usually get treated as “leadership” problems or “accountability” problems. That treatment is why I wrote my article Theory of Constraints Leadership – 2 Laws of Leadership. Leadership doesn’t need to be complicated, so I boiled it down to 2 laws. Law 1 is “Don’t be a sissy!” and Law 2 is “Just do it!” Pick what you’re going to do, own the results you get, then focus and get it done. You’re measured on 2 things. Reliability means you did what you said when you said you’d do it. Effectiveness means you didn’t waste time on things you shouldn’t be doing. I stand by the 2 Laws, and this article goes deeper into one specific place where owners get stuck, the shop that won’t follow the system. When owners hit that problem, here is what they usually try, and why each attempt isn’t enough on its own.

Assuming someone has it handled. I’m working with a shop right now where the executive handed the project to two of his people and disappeared. He doesn’t come to the meetings. He doesn’t ask questions. And they are not getting results. The defense sounds like this:

  • I’ve told them what to do.
  • I made it super clear.
  • They know what to do.
  • They know how to do it.
  • They know where to find me if they have a question.

Every sentence of that defense can be true. Not one sentence of that defense is “checking”. I call this pattern Responsibility Abdication. Handing your people the project is delegation, and delegation is right. Treating the handoff as the end of your part is abdication. You can delegate the authority, not the responsibility.

And the pattern is not new. Back in 2020 I was asked about an owner, call him Bob, who put a system in and then walked away from his business. I asked what questions Bob asked, what metrics he checked, and how often he checked them as he walked away. The answer was that he didn’t check anything, ever. So the issue was not that the system doesn’t work. It was a leadership and management issue.

Wishing and hoping. Wishing your people would just comply is not a strategy. W. Edwards Deming, the statistician and quality pioneer, studied a production line where management believed the workers were the problem. The data showed a stable system producing the defects, which meant the fix had to come from the system, not from the workers trying harder. His conclusion was blunt. “Any substantial improvement must come from action on the system, the responsibility of management. Wishing and pleading and begging the workers to do better was totally futile.”[3]

Telling them again. “Work the jobs in priority order,” said for the tenth time, is still just telling, and telling transfers knowledge. Knowledge is not the gap. Goldratt said it about respect. “To think that understanding it is enough to achieve it, for heaven’s sake, we are ignoring reality.”[1] In 2021 I posted, “Knowledge does NOT change behavior, what I need from you is a commitment to try what I ask.” Later in this article, I’ll show you what a commitment to try looks like.

A mission, vision, strategy. These are lovely words, but they can’t handle compliance for you. You can hang “On-time delivery is our number one priority” on the wall, and the floor will still run the easy job first.

Updating the employee handbook. Owners keep updating employee handbooks to drive the behavior they want. They end up with a pile of specific rules, because a rule gets added every time something bad happens. To be fair, measures and employee handbooks can be effective if done right, and later in this article I’ll show you an attendance policy that was done right. But a handbook that grows a new rule after every incident, like “Write up employees who don’t follow system rules,” changes nothing on the floor, because nobody is checking the floor.

Managing by checklist alone. A checklist can help. A full kit checklist, the list of everything a job needs before it gets released, is a helpful one. But managing by checklist alone doesn’t work, because a checklist tells people what to do, and knowing what to do is not the gap. A checklist also can’t find out whether it was followed, or why it wasn’t. I don’t believe giving somebody a checklist gets the results. I believe asking about the outcomes gets the results. The owner asks the managers about the outcomes. That question sends the managers to the floor to check that the system is being followed, and to ask why when it isn’t.

An incentive or bonus plan. An owner puts in a bonus plan for one or both of two reasons. The owner is thinking, I don’t want to have to play cop on the floor, and I want the bonus plan to do the work for me. Or the owner is thinking, the bonus plan is going to drive the results I want.

Then come the unintended consequences. People figure out the easiest path to the bonus, and the easiest path often isn’t following the system. So people do what earns the bonus. More units ship, but many of them aren’t the jobs customers are waiting on, and now customers are angry about late orders while the owner pays out bonuses. So the owner steps in and starts checking everything personally. Checking everything personally is hard to maintain, it eats the day, and the business isn’t getting run. So the owner tweaks the bonus plan and steps back again, and it all starts over. This is just one example of what can happen. There is also a right way to use a bonus plan, and I’ll show you that later in this article.

Playing cop. Another approach is to play cop, but to many owners that feels confrontational and icky. It also fails for a simpler reason. It’s hard to maintain over time. It’s hard to stay that resilient all the time, and eventually you take a vacation. We’re not looking for a cop who writes tickets. We’re looking for one who keeps showing up and asking questions.

ERP or scheduling software. ERP or scheduling software won’t get people to follow the system, because software can’t do the checking for you. It can’t ask the questions for you, and it can’t walk the floor for you.

Waiting for buy-in. If what you’re asking people to follow is counterintuitive, meaning it runs against what their instincts and habits tell them, then waiting for buy-in means waiting forever, because people cannot buy into something they can’t get their heads around until they’ve done it. If what you’re asking is not counterintuitive, then getting buy-in is great. I wrote a whole post on why you don’t need buy-in. Later in this article, I’ll show you how to start people on a counterintuitive system without waiting for buy-in.

Hiring better people. Nobody sets out NOT to hire the best people they can, and when the system isn’t being followed, it’s easy to conclude you need better ones. Hiring is important. If you hire people who can do the job, do it as well as they can, and are willing to improve their skills over time, you are in a much better situation. But hiring better people won’t get your system followed, because whoever you hire still has to learn how things work in your shop, and they learn that by seeing what others do and what’s important to their manager and leadership. If others aren’t following the system and management never asks about it, a new hire learns that it doesn’t matter, and the better they are, the faster they figure that out.

Some of these can help. None of them is enough. What’s missing is someone who keeps checking. To see why checking is the missing piece, and not just one more item on the list, look at why people drift away from systems in the first place.

Why It Keeps Happening

Think about what water does on a hillside. It takes the path of least resistance, and every time it runs, it wears its channel a little deeper, until diverting it is nearly impossible. People in a shop work the same way. They take the path of least resistance, and the longer they run it, the deeper it wears in. So if you want to know why your people don’t follow the system, look at what shapes their path. Three things shape it.

The measure is the slope. Water runs downhill, and people run toward however they’re being measured. If you have the wrong measure, you’re going to drive the wrong behavior. As Dr Eli Goldratt put it,

Tell me how you measure me, and I will tell you how I will behave.

Dr Eli Goldratt
Creator of Theory of Constraints

Years ago I was at a metal manufacturer, and I asked an operator how he knew he’d had a good day. His answer was, nobody yells at me. I asked what kept the yelling away, and he said that if he met the tons per hour budgeted for the day, he pretty much skated out of there and nobody said a word. So he ran the thick, easy product that made his tons. Notice that nothing in his good day connects to what the customer actually wants or needs or has ordered. The real constraint in that plant was the measure, tons per hour. And you don’t need a formal measure to have a slope. As my colleague Beau put it, even in the absence of a formal measurement system, “the owner’s questioning and reaction to responses becomes the measurement system.”

Here is what that looks like in a custom shop. Many of our clients start their jobs on a laser. A laser can be a $1 million machine, so they want to keep it busy, and they want to get the most out of the material. So the two measures are machine utilization and material utilization. To hit both, they look at their entire backlog, find the best nest, and run the laser in that order.

Neither of those measures is bad by itself. Here is what they cause:

  • Jobs get picked for the nest, not the due date. This means they use capacity to produce work that isn’t due yet and that they don’t have enough capacity to make what is due now.
  • The floor floods with work in process. Every nest starts a whole bunch of jobs, and they queue up behind the next operation.
  • Downstream departments play the same game. Once the parts are cut, each department gets to pick what’s easiest for it.
  • Due dates and dollars out the door both suffer. Jobs are running in the wrong order, and with that much work in process, less gets finished and shipped.

Nobody in those shops is trying to do anything wrong. “Keep the laser busy and get the best nest” pulls them there, the same way water runs downhill.

The old way is the worn channel. The second shaper is habit. People follow the path of how they’ve always done it, and they may believe their way is the right way. In manufacturing, the deepest channel is the efficiency habit. At a very young age we are taught to be efficient, and to be efficient we think we have to keep everyone busy all the time. Velocity Scheduling often runs counter to that habit, and running counter to a worn-in habit is a big part of what makes it counterintuitive. My 2 Laws article warns about this channel, “don’t let old patterns, past ways of doing things, and failures stop you”. James Clear describes the same pull in his book on habits. “Human behavior follows the Law of Least Effort. We will naturally gravitate toward the option that requires the least amount of work.”[4] His prescription for changing a habit is the same as mine for a shop, to “make the good habit the path of least resistance,” and I’ll show you how when we get to the Compliance Leadership Method.

Nobody checking means the water runs back. Suppose you divert the water into a new channel. The day you stop maintaining the diversion, the water starts finding its way back to the old bed, because the old bed is deeper. A system is the same. When a system that worked stops being followed, here are the explanations I hear:

  • The new people never bought in.
  • The old habits were too strong.
  • It only worked because of that one person.

I don’t buy any of those as the cause, because each of them only names the pull, and the pull is always there. What decides whether the pull wins is whether anybody is checking. Did they comply to the system? No, but why would they? Nobody was checking. Re-explaining won’t hold the water out of the old bed either, because explaining pushes against the channel instead of re-cutting it. Clients tell me they’d rather their people just comply, because the cause-and-effect thinking behind the system is not easy to grasp, and not easy to keep explaining. And while the explaining wears you out, the pull never lets up. Efficiency is so ingrained in us that at any opportunity, we tend to revert back to efficiency.

So why does the checking stop? Somehow, making sure people follow the system got a negative connotation. Some owners see it as confrontation and want no part of that, so the job of making sure the system is followed doesn’t get done. Goldratt warned about that instinct. “When conflicts are not resolved but instead are pushed under the carpet they do not disappear. They grow.”[5] And for managers, I have a two-part hypothesis about why they stay off the floor. I suspect a portion of them are intimidated. They think they’re supposed to know more than the operators, and they can’t run the machine, so they don’t want to leave the office. Others came up as machinists, aren’t intimidated by the machines or by knowing less than the operators, and simply don’t like confrontation. Both groups stay off the floor for different reasons, and the answer to both is the same. Asking isn’t confronting, and I’ll show you exactly what the asking sounds like when we get to running the system.

So when your people aren’t following the system, where do you look first? It depends on where the shop is. If a shop hasn’t put a real system in yet, suspect a bad measure, a false assumption, or a policy first. If the system is in and was working, suspect drift or resistance. Either way, start the diagnosis with the slope and the channel, not with the people. Goldratt said the same about the conflicts in a company. “The conflict exists not because of a person. Never because of a person. It’s inherent in the situation.”[2] Defiance is real, and I’ll show you how to sort it out later, but it is the last thing to check, not the first.

Before you decide your people are defying you, check whether they’re just running downhill.

That picture is also why the typical solutions fall short:

  • A poster doesn’t change the slope.
  • A bonus plan tilts the slope the wrong way.
  • A handbook rule doesn’t re-cut the channel.

What changes behavior is a system with the right measures, worn in by someone who keeps checking. That is where Compliance Leadership starts.

The Core of Compliance Leadership

Compliance Leadership comes down to this.

  • The system is measured on results.
  • The people are measured on compliance to the system.
  • Managers are responsible for designing and implementing the system, and therefore responsible for the results.

The people measured on compliance are the people on the floor. Upper-level managers can and should be evaluated on outcomes, and judgment roles, like schedulers and engineers, are hired and measured on judgment. I’ll come back to why the levels split that way.

Start with the system. In a shop, the result the system is measured on is flow. Flow means getting more jobs done with the same people and resources, increasing productivity while still balancing due date performance. The owner or the operations manager owns flow. It gets reported to everybody, and it is not scored per person or per team, because no one person controls it. The system produces it, so the system carries the measure.

The people carry a different measure. Your employees are responsible for compliance, for following the system. They follow the policies, follow the procedures, and work to the measures. And that split is what makes the whole thing fair. If the system is followed and the results are still wrong, the problem is the system, and the system belongs to management.

Managers are responsible for designing and implementing the system, so the results belong to them too. This is not a new position for me. In 2009 I wrote that our policies, procedures and measures are perfectly getting us the results we are currently getting.

And the thinking behind it goes back to Dr Eli Goldratt, the creator of Theory of Constraints, who I learned from and worked with. Goldratt didn’t really talk about leadership by name. But he said everything about measures and behavior in one line, and his book prints the whole statement in capitals. “TELL ME HOW YOU MEASURE ME, AND I WILL TELL YOU HOW I WILL BEHAVE. IF YOU MEASURE ME IN AN ILLOGICAL WAY … DO NOT COMPLAIN ABOUT ILLOGICAL BEHAVIOR.”[6] Behavior follows the measures, the measures are part of the system, and managers design the system. So the results sit with the managers.

Goldratt is not the only one who puts the results on management’s side of the table. W. Edwards Deming estimated that most troubles and most possibilities for improvement “belong to the system (responsibility of management)”. He put the number at 94%.[3]

Who “The People” Are

The level decides what a person can actually move. That is why the compliance measure belongs to the floor, outcomes to upper-level managers, and judgment to the judgment roles.

Upper-level managers
evaluated on outcomes
Judgment roles, the schedulers and the engineers
measured on judgment
Floor roles
measured on compliance to the system

Floor roles are measured on compliance to the system. Understanding why the system works is optional for them, as long as they follow it. Knowing the outcome is not. That’s how they can tell when it’s about to miss. And holding operators to compliance does not mean the operators are not valued. You can both hold your operators to system compliance and value their input.

Upper-level managers are different, because they can see and move the results. They can and should be evaluated on outcomes.

Judgment roles sit in between, the schedulers and the engineers. They are hired and measured on judgment. Their judgment won’t always be perfect, and that is exactly why the method they use matters too.

So why not hold everybody to the outcome? I would love to. I would love to delegate the outcome at every level, to hold everybody to the appropriate outcome at whatever level they are. But what we do is counterintuitive, and nobody on the floor would figure the method out on their own, so handing the floor the outcome alone can’t produce it. The method has to be given, and in Velocity Scheduling, and in Theory of Constraints in general, the method is as important as the outcome. It doesn’t work to delegate the outcome to the floor. We have to have compliance at the floor. Experience doesn’t settle it either. At one shop, an operator who had been there for years drilled a hole into a press that should never have been drilled, and it was destroying the machine. Ask the floor how to ship more this week, and the answer is most likely wrong. That isn’t because the people are bad. It’s because the right answer is counterintuitive.

Goldratt wasn’t asking for that either. “I am not saying that now we have to empower everybody to do everything,” he said, and he pointed out that not all people want more authority.[1]

The test behind that answer is line of sight. An outcome belongs in a person’s role only if that person can touch or influence it. It is wrong to evaluate people on an outcome they don’t know how to achieve and can’t directly affect. Goldratt put it this way. “We shouldn’t blame a person for something that he cannot do a thing about it.”[1] You need to build the system and the measurements to achieve the outcome.

Here is what it looks like when that test gets skipped. Beau recently talked with a man who owns five Jimmy John’s sandwich shops. Each store is measured on prime cost: the labor and the materials that go into a sandwich, the meat, bread and cheese, divided by sales. The goal is to keep that percentage under a target.

The intent makes sense. If costs stay a small enough share of sales, the company keeps more of every dollar it takes in. That part is true. The owner expected better profit from every store.

Now look at it from the store manager’s side. There are only two ways to bring a percentage down. You raise the bottom number, sales, or you cut the top number, costs. The manager can’t control sales. He’s running a store, not the marketing. So the path of least resistance is labor. Beau put himself in the store manager’s shoes and told the owner what he’d do: “I’m going to fire people as soon as I can.” Instead of scheduling someone from 11 to 2, the manager schedules them from 11:45 to 1:15, a shift so short it barely pays for the Uber ride to work.

Here’s what happened. People quit. For about three weeks, the manager hits every number and the owner likes him. Then the store is short-handed and in chaos. When Beau described it, the owner said, “that’s exactly what happened.”

The impact is the opposite of what the measure was for. It was meant to protect profit, and it destroyed some of it, because the only lever the manager could reach was cutting the people who make and sell the sandwiches. If a measure can do that in a sandwich shop, it can do it in yours.

The Great Game of Business takes the other side of the question of whether to hold everybody to the outcome. It gives shop floor workers the business metrics and expects them to move those numbers. I would love for that to work, but I don’t think most shop floor workers can, and the reason is the same counterintuitive problem. When the right way to run a shop runs against intuition, the causal understanding people carry around is typically wrong. They don’t understand how what they do today moves a business metric. That isn’t a knock on the floor. Even the managers I coach find it hard to get all the causal relationships in 14 weeks of coaching.

None of this means we keep the why from them. Goldratt said that it is “as important to tell them why” as it is to tell them what to do.[7] I agree. He also pointed out when we usually give it. “When the lieutenant messes up, only then do we explain the why. We call it on-the-job training.”[7] The why should be part of the instruction when you show someone how to do it. We explain the outcome, the why, and we give them the process, the what to do. Some people perform better when they understand the why. Some will want to understand it, and some won’t. They don’t have to understand it to do it. The ones who want to are what I call the keepers, and I’ll come back to them.

We so often give our people the what and why, and then assume that they’re going to follow it forever.

The Manager’s Job Has Two Halves

The manager’s job is to 1) build and improve the system and 2) make sure it’s followed. Making sure the system is followed is the half of a manager’s job that seemingly few want to do or know how to do. I hear a lot of talk about accountability and see little of it in practice. I suspect that most owners, managers, and supervisors don’t know how to hold their team accountable before the miss happens. The rest of this article is about the how: getting the system right, checking while the job can still be saved, and taking responsibility for the results yourself. The Compliance Leadership Method at the end lays it out step by step.

But before the how, you have to accept that the second half is really yours, and my definition of leadership says why. Leadership is making something happen that would not have happened on its own. As supervisors and managers, we create systems and processes and ensure compliance with those systems and processes. Compliant systems and processes don’t happen on their own, and most of the time, neither does compliance, particularly when we’re trying to implement something counterintuitive.

And I’m not the only one who hangs the job on the manager. Elliott Jaques, the organizational psychologist, defined a manager as “a person who is held accountable for the outputs of others, for sustaining a team capable of producing those outputs, and for giving effective leadership to that team.”[8] He puts the accountability for the output on the manager, and that is exactly where I put it. The operator’s responsibility is to follow the system. The manager’s responsibility is to make sure he does.

How to Hold Employees Accountable: Check That the System Is Followed

So what does making sure it’s followed look like? Telling them, even telling them repeatedly, is not enough. There have to be checks, because you can’t abdicate your half of the job and trust that the system runs itself. Trust but verify. This is not Field of Dreams, where just because the system is built, they will follow it.

Checking is not hovering over people.

Checking is two actions.
1You ask questions.
2You go look.

Start with the asking. When I’m in a shop, I go out on the floor and ask questions. Oh, that’s really interesting. Now, why do you do that? People are usually happy to tell me and educate me. And the questions themselves hold people to the system, because a question about the procedure, asked again and again, tells everybody the procedure matters. Holding people to the system is firm but not confrontational. You stay on top of it, you’re firm, and you continually check on it until it becomes the new habit. It’s asking, not policing. The cure to confrontation is curiosity. Try “How’d you know to do that?” or “What made you want to do it that way?” And look for obstacles, not bad decisions.

It’s the checking that demonstrates something is important and that non-compliance is not an option. If nobody’s checking, if nobody’s paying attention to something you’ve been told to do, do you keep doing it?

Larry Bossidy, the former CEO of AlliedSignal and Honeywell, quotes Dick Brown, who ran EDS, on exactly this.

Leaders get the behavior they exhibit and tolerate.

Dick Brown, who ran EDS, quoted by Larry Bossidy [9]

Here is how fast the behavior follows the checking. There is a story about a father who let his teenage daughter use the family car as long as she washed it every Saturday. He checked every Saturday, and it worked. Then he was out of town two Saturdays in a row. When he got back the car was dirty, and she said, “Oh, are we still doing that?”[10] Two missed checks, and the agreement was gone. If that happens with one teenager and one car, think about a whole shop. In the Velocity Scheduling System, the father’s Saturday check has a shop version, the 5-minute daily meeting we run at the board. Miss it twice and you will hear the shop’s version of “are we still doing that?”

The other half of checking is going to look. At one manufacturer we work with, we asked whether they had full kit for the next several jobs. The answer was, “We should.” We told them, “Should and do are two different words.” Later on that same call we heard why that matters. A job had been released because the inventory system showed the motors for it were in stock. The motors could not be located. The count showed the motors were there, and the count was wrong. A count is not a check, so the Velocity Scheduling System has someone physically check that the full kit is there before a job is released. When the question is whether something got done, ask to see it, not to hear about it.

The Checking Chain

So who does the checking? It runs on a chain with two links.

Managers and supervisors do the asking on the floor. They ask about the policies, procedures, and measures of the system, because those drive behavior and keep people following the system. The owner asks the managers about results, what moved, what slipped, and why. The owner watches results-level effects, and the managers and supervisors manage the causes.

Both links have to hold. There has to be compliance from the manager to the shop, but also from the owner to the manager, and I think neither of those is happening in most shops.

Owner
asks the managers about results
Managers and supervisors
ask on the floor about the policies, procedures, and measures
The floor
follows the system
Responsibility goes up.

And nobody is too senior for their link. Andy Wilson, a retired Walmart executive who walked stores with Sam Walton, tells a story. “Sam walked into the store with a tape recorder, and he would record any out-of-stock he saw.” “Those recordings would be transcribed that evening, and he expected an answer the following day.”[11] When Sam found an empty shelf, he would likely call for the department manager and try to get to the root of why the shelf was empty. Sam Walton founded Walmart, and he was still checking shelves himself. If Sam Walton wasn’t too senior to check, nobody is.

When the check turns up someone who isn’t following the system, who has to act? Your employees are responsible for compliance, for following the system. You’re responsible for the outcome, because you’re responsible for the system. And if your people are not complying, that is also your responsibility. Your job is to find out why and to help them understand how and why to comply. Their responsibility doesn’t go away. Yours kicks in. A miss in the results is not the employees’ mistake. It’s the people above them. And it keeps going up the chain. If a manager isn’t checking, it’s their manager’s fault, then the manager above that. Responsibility goes up.

Compliance Includes Speaking Up

The chain runs top down, but compliance itself is not silent obedience, because there are moments when following the steps isn’t enough. People are accountable for speaking up and following the system. And there are two kinds of speaking up, which are not the same thing.

The first kind is offering suggestions and ideas, and this kind is voluntary. Suggestions are never required, and they are never punished. But they come with a duty on your side, because if people give you suggestions and you do nothing with them, they won’t continue to give you suggestions. Goldratt put it more bluntly. “Ignoring someone’s idea is a rude form of disrespect.” And “The key to managing people is mutual respect.”[1] So circle back on every one. If a suggestion turns out not to be helpful, you still have the responsibility to circle back and explain:

  • why you didn’t do it,
  • how you may have modified it,
  • what you did do, or
  • how you used it.

Sometimes people give suggestions that aren’t helpful because they don’t fully understand the situation, and that in itself is telling, because then there’s an opportunity to further explain.

When you have a concern about someone’s idea, start by saying what the idea would fix. Goldratt said “The best way to praise an idea is by summarizing its resulting benefits,” because until then the person isn’t ready to listen. “Your idea is very nice,” followed by a but, is what he called plastic politeness.[1]

Goldratt called a concern about an idea a “yes, but.” Typically it’s a small yes and a big but. In Theory of Constraints, the formal name is a negative branch reservation. His advice was that “The best way to prepare for criticism is to understand the detailed logic that connects the idea to the probable negative ramification.”[1] A lot of managers and owners struggle with that logic side, and you don’t have to draw a logic tree to do this. Ask two questions. What part of this idea worries you? What negative do you think could happen? Then work out together whether the person misunderstood the idea or is right. If the person is right, ask what you could change to prevent it, and then be quiet. Goldratt said every “yes, but” contains a gift, and to “control yourself for about 10 seconds,” because the person who raised the concern usually has the intuition to come up with the fix.[2] You end up with a better idea, and the person with the “yes, but” helped build it.

The second kind is not voluntary. When someone is following the procedure to the letter and can see the standard is about to miss, speaking up is part of following the system, and staying silent is the passive-aggressive behavior that is not okay. Goldratt said the same about project managers, who have the responsibility but almost none of the authority. Their job is to raise the flag. “We can blame a project manager for not raising the flag.”[1] Here is an example. At one manufacturer, a line was running at the rate it had been given, to the letter, and that rate was about half of what the line could really make. Everybody on the line knew it, and the rate stayed where it was. And they must have been doing this well before we showed up. When we suspect the outcome will miss, blindly following the rules without question is not okay.

Now, some people will stay silent anyway, because they bring their baggage with them from bosses who punished speaking up. Some will even use the George Costanza method of looking busy to avoid talking to a boss. You can’t fix that with an announcement. You have to show them, over time, that suggestions get circled back on and that speaking up gets thanked. It has to be demonstrated through action. And probably still not everybody comes around. This isn’t an all-or-none thing.

When Employees Aren’t Following Procedures

Sooner or later the checking turns up a workaround, and what you do next decides whether the system gets better or the person gets blamed. Find out why first, and have the conversation, because a workaround is typically one of three things:

  • It is defiance.
  • It is a sign the system needs fixing.
  • It is a lack of understanding of what the system is asking.

In my experience, most workarounds turn out to be people who didn’t understand what the system was asking. They probably heard the instructions over and over, but they didn’t truly understand them, so they went with what they thought they understood. Like water on a hillside, they followed the path of least resistance. But you still sort each workaround into one of the three, because some are defiance, and some signal a needed improvement or a hidden false assumption, and those are the ones that make the system better. Workarounds are not an all-or-none thing. We have a logical thinking-process tool that helps us sort which of the three we’re looking at, so if you have a workaround you can’t sort, reach out and we can help you process it.

The sort matters because each answer gets a different response. A sign the system needs fixing becomes a system change. A lack of understanding becomes a learning opportunity. Either way, circle back and tell the person what happened. And if the SOP itself was wrong and was followed to the letter, the issue is the SOP. It’s still a system problem.

Goldratt gave a good test for one kind of system problem. When someone keeps coming to you with a fire, he said to look for a mismatch between that person’s responsibility and authority. In his example, the person in charge of shipping was responsible for on-time shipments but couldn’t call the client about a missing ship-to location, because only the account manager was allowed to call.[7] Nobody had written that rule down. The system was blocking the job. He said the average person is under about seven of these misalignments, and that bosses are mostly blind to the ones below them. And he said not to fix it alone. Find the correction together with the person, so it’s their idea.[1]

Money is a common version of this. People are told spending is bad, so they keep doing something unproductive instead of asking. Someone may still be hauling a 100 lb nitrogen cylinder around to charge the accumulators, when a 20 lb cylinder they could tow themselves would be faster. They don’t have the authority to buy it, and they don’t want to suggest more cost.

Defiance is the last case, and it has its own sequence:

  1. Can they do?
  2. Will they do?
  3. Is there a fit?

If you find someone who can do (they understand and have the skill) but won’t do, now you have a decision to make, whether to keep working on the why or to replace that person with someone who will. The same three questions guide my hiring, but their importance runs in a different order:

  1. Will do matters most, because people can be trained.
  2. Can do is nice, it’s a bonus.
  3. Fit comes last. We obviously want people to fit with our system, our people, and our culture, but most of the culture is defined by your system.

Drift Is a Checking Failure

When a system that worked stops being followed, a little at a time, I call that drift. In my experience, VSS sticks for most shops. When it doesn’t, it usually followed a personnel change. The champion left, or got promoted. Sometimes the whole shop got slammed for a month. But the champion leaving is not a separate cause. It is just the most common way the checking stops. Goldratt saw the same collapse. “Even if the implementation does work, if you don’t change the bureaucracy to fit the new rules, these people will be promoted, these people will leave, and then everything collapses.”[2] Asking about a new practice is what holds it in place against the pull of the old way, whether the practice is safety, a procedure, or a Velocity Board.

Drift is a checking failure.

Owners ask me, so do I have to check forever? My answer is built into how I set up the checking. Checking is part of the job, so it can’t depend on one person’s memory or willpower. The checking belongs to a role, and when that person leaves, the next person in the role takes it over. The other half of my answer is to build the shop so the right way is the easy way, which is part 2 of the Compliance Leadership Method at the end of this article. Then managers can check less often, every week instead of every day, but they never stop checking.

And does it matter that the owner at the top keeps checking too, not just the managers? It does, and the difference is measurable, and not just in my experience. A study of 70 companies that put the same kind of management system in place found productivity gains averaging 56% where top management was highly committed, and 6% where it wasn’t.[12] Same system. The difference was the people at the top. In practice, that commitment looks like the owner asking the managers, and never stopping.

Don’t Ask for Buy-In. Ask If They’re Willing to Try.

The manager carries the compliance half of the job, and that half starts on day one. So how do you start a counterintuitive system without waiting for buy-in? I don’t ask for or expect buy-in at the outset, because getting buy-in ahead of time is very difficult when the solution is so counterintuitive. People cannot buy into something they don’t understand, and all the agreeing up front is just talk. We look at the logic together, they agree it makes sense, and at best they can honestly say it looks good on paper.

Here is where I part ways with Goldratt. He was fond of saying that people don’t resist change, they judge it, and that the perception of risk dominates the judgment.[13] I agree. The standard Theory of Constraints answer is to get buy-in up front, agreeing on the problem, the direction, and the solution, one layer at a time.[14] His process for it was built for top management, about 15 people off-site for four days with a facilitator, and then the buy-in filters down.[2] The person on the floor wasn’t in that room. I agree with explaining the logic. But in my experience, people can give you the logic back and still not buy in, and waiting for buy-in slows down the entire project.

I was on a coaching call recently with the person running one line in a shop. She had fewer people available and a higher target, so she increased her work in process, which is the exact opposite of what she should have done. We asked her some questions. If you add work and nobody is free to work on it, what happens to productivity? It goes down. If things flow faster, is it easier or harder to make a higher output target? Easier. She gave the right answers, and she still wanted to increase work in process. “It just makes sense to me,” she said. A few minutes later she told us, “Sometimes it’s just hard. The mindset is hard.” That is what counterintuitive does. Goldratt came to see some of this himself. Later on, he wrote that when people have no experience with a change, “explanations are not sufficient,” and that the next step should be a test.[15] But he still wanted the explanation first. I give the why too, and I don’t wait for them to agree with it. I ask them to try it.

So the gate I use is one question.

Are you willing to give it a try? Any answer other than yes is a no.

Really, I’ve learned that the hard way. Goldratt ended his own team process with the same question. After a team worked through every obstacle, he asked, “should we try?” The answer he was after was “let’s give it a shot,” and he said, “That’s what you need, nothing more.”[1] I just ask it first. Then, after the first result, our job as coaches and managers is to point out what they did, the improvement that resulted, and why. The result is what converts into belief, and that belief creates momentum to keep implementing.

The 2 Laws article says the same thing about the person who resists. “Ask them if they are willing to try (really try) it. If so, then get on with it. If not, it may be time to replace that person with someone who is.” And the try has to run long enough to be a real test, because, as I wrote there, “people either 1) don’t try something long enough to see if it’s going to work – they get distracted by the next new shiny thing; or 2) they stick with it forever despite the poor results.” One of the best leaders I’ve worked with was the client in that article whose “team was struggling with buying into VSS but he held firm (definitely not a sissy) and then focused on doing it. And doing it right, not some watered down version.”

Paint the Outcome. Give the Method When It Matters.

What you hand people to follow is not a task list, because you can’t list every task. The environment changes, the situation changes. It’s too dynamic. The closest you can come to giving work instructions is to paint a picture of the outcome you’re trying to achieve. What does good look like? What does the end state look like? Then provide as few specifics as possible, only what’s necessary and sufficient to achieve the end state. Goldratt saw the same problem with piling on detail. “The more detailed the instructions, the less the empowerment.”[7]

So when does the method itself get specified? When getting it wrong would be a disaster, but that is not the only time. I also require the method when it’s so counterintuitive that nobody would figure it out on their own. Velocity Scheduling is exactly that counterintuitive case. You cannot simply tell people, “Increase flow.” They don’t know how to do it, and you have to give them the method.

Owners push back here with two good questions. The first is, won’t this turn into a rule book? No, because a rule book is what you get from a different process. Rules get added every time something bad happens. A car accident happens at a dangerous intersection, so they take out the stop signs and put in lights. Something bad happens to somebody, and new rules get put in place. You end up with a Dorian Gray of employee handbooks that try to address everything that has gone wrong and could go wrong. That is not my approach. That’s not what systems mean. Systems, good systems, should work to outcomes.

Business researcher Jim Collins watched the same pile-up and traced it to people instead. “Most companies build their bureaucratic rules to manage the small percentage of wrong people on the bus, which in turn drives away the right people on the bus, which then increases the percentage of wrong people on the bus.”[16] The same principle applies to scheduling, and there I teach it bluntly. “We do NOT detail schedule (STOP adding detail).”[17] Having a more detailed schedule just means we’re more wrong. A thicker handbook is to people what a more detailed schedule is to jobs. More detail just means you’re more wrong.

The second question comes from owners who read the Toyota literature. Toyota has operators improve the standard, so why not us? Toyota is a repeat production environment. Toyota sets up to run a model vehicle, and that model vehicle may run for 5 years. I’m dealing with custom manufacturing environments where the jobs differ, so operators improving the standard themselves is a little more difficult. We want to engage them on setup reduction, how we could get this set up faster, and on their observations of how to do things better. But I stand by this. The manager is responsible for improving the system, because we’re dealing with custom manufacturers, and we’re dealing with implementing a counterintuitive, counterhabit system.

In my system, the manager owns the standard. People on the floor still have a job, which is to speak up when they see the standard is about to miss. The keepers, the operators who want to understand why, are the people I watch, because they can move into a role where improving the system becomes part of their job. I’ll come back to the keepers at the end.

The Bonus Plan, and the Right Way to Use One

We don’t believe in incentives for the shop floor. We believe policies, procedures, and measures drive behavior, and incentives are really hard to get right, and hard to tie to something people can clearly and knowingly impact. They become entitlements, and they break the inherent employer-employee relationship, which is that I pay you for your best work every day. If I pay you for your best work, then what is a bonus paying for?

A bonus pays for a number.

Being against incentive programs never meant doing nothing. In the 2 Laws article I tell about a client who wanted to do “an incentive program (which generally I’m against)” and I told them I didn’t care if they wanted to bring clowns in on Fridays – but to DO SOMETHING.

So why not do something with a bonus plan? Because a bonus plan can backfire. Measures drive behavior, and a bonus is a measure with money on it, so it drives the behavior harder. You have to balance productivity with due date performance. If you want to weight productivity more heavily, you can, that’s a business decision. But you better be clear on what you want, because the measures will drive it, and if you put a bonus on top of those measures, for sure it will drive it.

There is a right place for a bonus plan, and it is defined by responsibility, not title. You have to be senior enough that you are responsible for the systems and the output, and the output has to be directly tied to whatever measures you’re going to bonus on. In small shops, that’s not hourly people. The test is line of sight. If you don’t have any idea whether or not you’re going to get the bonus, that’s a clear indication that you don’t have control over it.

What does it look like when the bonus sits on a target handed down from the top instead? Jeffrey Liker, the University of Michigan professor who has studied Toyota since the 1980s, watched Ford adopt Toyota’s policy deployment. Set the policy at the top, mainly profits in Ford’s case, deploy it down through the levels with some give-and-take, and once everyone’s targets are aligned, let it rip. “Those who met their targets were rewarded, and those who didn’t . . . well.”[18] The thinking at Toyota, he found, was night and day different, because in Toyota’s version the leaders stay connected to the details down at the level where the work happens, instead of issuing targets and reviewing results from above.

I lived the Ford version personally. I was a senior manager at Anheuser-Busch, and we used KPIs. There were company KPIs that might have been related to driving profitability, but those KPIs get brought down by the levels, and by the time one reached my level, there was really not a direct path back to wherever the KPI originally started. We really didn’t know how the decisions we made day by day impacted the overall company. The incentive was to try to move whatever your KPIs were. Some of them weren’t even quantitative, because it’s really hard to take something you don’t touch and can’t directly influence and drive it down an organization. They end up distorted. That experience is the reason for my rule. Bonus only the people responsible for the systems and the output, with the measure directly tied to that output.

The distortion even has a name, Campbell’s law, from the social scientist Donald Campbell. “The more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor.”[19] The job shop version of the same distortion is each department’s efficiency or utilization number, which looks good locally while it fights the due dates.

Owners also ask the fairness question. Doesn’t measuring on results reward the wrong people? Larry Bossidy tells a story about exactly that.[9] Dave and Mike made their numbers last year, and Sue missed hers, so a mechanical evaluation pays Dave and Mike and skips Sue. Look closer. Dave coasted to success on a stronger-than-expected market, and if he’d been doing his job well, he would have beaten his projections by 20%. Sue missed because a raw material shortage unexpectedly raised her costs by 20%, and her results would have been considerably worse if she hadn’t quickly accelerated some planned productivity improvements. Mike brought in his earnings by halting the development of two new products and forcing product into the distribution pipeline, borrowing from the future to make his numbers today. Bossidy’s verdict is mine. “If anybody should get a bonus, it’s actually Sue.” Because “When the wrong people get rewarded, the whole organization loses.”

Here is my read on the three:

  • Dave rode a wave, not the system, and in my shop he would still be getting feedback that he’s not following the system.
  • Mike pulled forward the jobs that make more money and went late on the others. That’s a business decision he’s allowed to make, but the system would point out that it’s not ideal if you want to be productive and on time.
  • Sue is the one I’d go sit down with. Why don’t we behave like that all the time? What did you do? How did you do it? The Velocity Scheduling System is how you gain and sustain that higher-level productivity consistently, not as a fire drill.

Underneath all three reads is the same rule. Taking the process into account matters, and somebody who jumped to their favorite solution gets more heavy-handed feedback from me than somebody who followed a good process and got a bad result.

That leaves the hourly people. If a bonus is the wrong tool for them, the right tool is pay. Our view is, instead of bonuses for hourly people, make them the best paid in the area, with the right measures in place and a good visual system. Best paid means best paid for what they do within your area, and how big that area is depends on where you’re located. That could be a 200-mile radius. The idea is that if you have an opening, everybody wants to work there, and you have lots of people to choose from. Skilled machinists are still scarce, but replacing someone who isn’t following the system is easier if you’re paying the best wages around. A high-performing environment like that also tends to create more peer pressure to perform well.

I married into a job shop, and that shop came with proof that replacing a bonus with pay can be done right. Before we met, in the 1990s in Denver, my husband Brad owned the shop, and he had a bonus plan that became an entitlement that really wasn’t effective and wasn’t getting results. To get rid of an entitlement, he needed a really good replacement. So he replaced it with a 25-cent-per-hour raise every quarter, and he did it for years, as long as the company could afford it. His objective was for his people to be the best paid for what they did within 1,000 miles.

What if you can’t afford the raises it takes to make your people the best paid around? That may be true to start. With traditional bonus systems, the measures often aren’t correct and aren’t tied to the bottom line, so you end up paying a bonus when you’re not making more. Repurposing that money is one approach. And if you’re paying a bonus to keep your best machinist because they’re hard to replace, look at your systems. A system with the correct measures, one that actually gets more dollars shipped out the door, should produce more profit to provide raises. If your systems aren’t producing more out the door and sustainable profitability, you’re not using the right systems.

Alongside the raise, Brad dealt with a second problem in the same shop, attendance. Attendance looks like an HR issue, but it is a system compliance issue, because people are your capacity. If people don’t show up, you can’t run your system. Brad put it this way. “I came around to understanding that I was paying for availability, but I needed to manage for productivity.” People were showing up late, leaving early, or not showing up. So vacation had to be earned. You earned vacation by being there and on time, and being late or absent consumed it. You had to be there enough to earn the vacation. It is a very unique and different approach, and it substantially improved attendance. The attendance problem went away. In Brad’s words, “This is not a popular program, but nobody fought me on it either. It accomplished what its purpose was.”

And attendance feeds straight back into scheduling. In VSS we do not detail schedule, but we predict when every job will start and finish, and those starts and finishes depend on people being there and on time.

I have the full details of Brad’s employee availability solution written up, including how the earning bands worked and how to adapt it to your shop. Download the full employee availability solution and you’ll get the design we used.

Compliance Leadership in the Age of AI

So where does AI leave all of this? AI changes the kind of compliance. It never removes it. There will always be a little bit of that compliance gap, no matter how much of the work the tools take over. Software can pick the job to run next, but that doesn’t mean the operator out on the floor is going to pull that job and run it. He says, oh no, I don’t want to run that one, I’m on this setup, I want to go run this other one. You still have the compliance issue. However fast a tool produces a schedule, somebody still has to check that the floor runs it.

Trust but verify also applies to the tools themselves. Robert Reed, an industrial maintenance manager, said it about checking AI agents. “On the floor nobody signs off a work order because the tech said it’s fixed. You watch it run.”[20] The full kit check is the manual version of that stance, because a count can say the parts are there when they aren’t.

And AI raises the stakes on one thing in particular. AI can output so much information that the limit becomes the human ability to absorb it, make the decisions, and judge. I wrote a full article on that capability, Critical Systemic Judgment. Compliance Leadership protects judgment in two ways. First, the checking belongs to a role, so however much information pours in, asking why stays somebody’s job, and it keeps happening. Second, the daily asking is how you find the curious people, the ones who ask why, and they are the ones who can grow into the judgment roles. The last part of the Compliance Leadership Method is about them.

The Compliance Leadership Method

Start with the vision. In a shop running Compliance Leadership, the system gets the results and keeps getting better. People follow the system and speak up when it’s going to miss, and the ones who want to give feedback feel heard. Leaders don’t disappear after the handoff. They delegate the authority and keep the responsibility. Managers walk the floor and ask questions, what’s going on with this, how’s following that procedure going, and the owner asks the managers about results. Holding people to the system feels like a normal conversation, not a confrontation, because it’s asking, not policing. And when something isn’t followed, responsibility goes up, not down.

Water is going to run downhill no matter what you do. Divert it correctly, and it generates power. Your people are the same. Their effort is going to go somewhere. Put them on the right path, keep them there, and that effort turns into results.

Here is how it all ties together. Six parts get you there.

1. Get the system right

The design principles are simple:

  • The system is measured on results and the people on compliance.
  • Measures drive behavior, so check that yours aren’t pulling people the wrong way.
  • Specify work by end state, add how to do the work where that matters, and stop at necessary and sufficient, because good systems aren’t rule books.

Goldratt tied this to empowerment. “The quality of the procedures that we are using determine whether or not we can empower.” And “If we don’t put logical procedures inside the company, empowerment becomes a joke.”[1]

We all have the best intentions with our measures, but people will follow them down the path of least resistance. In Theory of Constraints, we test a measure before we put it in place with a negative branch reservation. If we measure this, what could go wrong? How could someone hit this number and hurt the business? Do the same with any incentive before you put it in place. What’s the fastest way to get the reward, and how could it be misused? Whatever that answer is, that’s what you’ll get.

The specific policies, procedures, and measures come from the system you run. For scheduling, VSS gives you the policies, procedures, and measures. Compliance Leadership is how you make sure they get followed. Some of the standing scheduling rules are public, and I published them years ago. “Reduce WIP.” “Green/Yellow/Red is our one and only priority system. The Velocity Board communicates what to do.” “We do NOT detail schedule (STOP adding detail).” “We focus on finishing what we have started rather than starting more stuff.”[17]

2. Make the right way the easy way

People take the path of least resistance, so make the system the path of least resistance. You do that with tools, fixtures, and visual systems. What can we do so that an error is not even possible? Is there a tool? Is there a fixture? Think of a guardrail. Nobody trains a driver to stay on the road. They bolt steel to the edges so it holds whether they’re paying attention or not. Training teaches people the right answer. Architecture makes the wrong answer impossible. The architecture doesn’t replace the checking. It lets managers check less often.

For scheduling, the visual system is the Velocity Board. At a glance, operators know what to work on next, managers know whether jobs are on schedule or behind, and jobs worked out of priority order show. Gone are the days of one person scheduling in their head, because one person’s head is not a system.

3. Start it

For a counterintuitive system, don’t ask for buy-in up front. Ask one question. Are you willing to really try it, say, running the board for 30 days? Then, while they try it, ask for their input and their experience:

  • How can we do this better?
  • What did you experience as you tried it?
  • What did you notice?

Take the feedback into account and feed it back, telling them whether you did something with it and why. People commit more to their own ideas, and for something counterintuitive, trying it is how it becomes theirs. Then point out the results and the why behind them, because the results are what convert into belief. Trust is earned by action over time.

4. Run it

Your people try it, they get the results, they start to buy in, but then you have to keep reinforcing it, because they’ll forget. Our habits are so ingrained. Everyone was taught efficiency for so long that they will go back to the efficiency mentality if the management doesn’t stay on top of it. And the moment you stop matters, because as soon as you stop asking, as soon as you stop probing, they think it’s not important anymore. So the cadence can slow down, from daily to weekly, but it never stops, and because the checking belongs to a role, it survives a personnel change.

What do you ask? Any of these is checking.

  • When you see a pile of stuff, ask what’s going on with it.
  • If you’re implementing a new initiative or making a change to your system, ask people about one of its policies.
  • Ask how following a procedure is going.
  • Ask about a measure. What is it this week, and why?

For scheduling, the daily ask is one standing policy. Are we working jobs in priority order? Is the full kit there before a job is released?

Why questions? Goldratt’s tools for logical thinking are powerful, and a lot of managers and owners find the logic and the trees intimidating. These questions do much of the same work without the trees, and Goldratt’s own process was heavily based on the Socratic approach.[2] We are all meaning-making machines. Give me a little detail, and I’ll fill in the rest with my own assumptions. Asking why someone runs it that way brings those assumptions out, and it’s much better than one more statement from you.

If your managers avoid the floor because they think they need to know more than the operators, here is the answer. Checking doesn’t take technical proficiency. Managers think it does, and it doesn’t. Ask the operator why he runs it that way, and he’s usually happy to tell you and educate you, and the answer you get back is the check. To me, it’s always worth doing. The balance to strike is the one from the 2 Laws. “‘Just do it’ does NOT mean steam rolling your people (but at the same time don’t let THEM be sissies!)”

What about putting the numbers up on a big screen? Visibility is not checking. Raja Shembekar, a Denso vice president, toured supposed Industry 4.0 benchmark plants and found “a lot of fanfare and wonderful displays of data, but almost no action to solve real problems.”[18] He has a name for screens nobody acts on, electronic wallpaper. The late-job report on the big screen that nobody acts on is electronic wallpaper too. Business researcher Jim Collins found that the good-to-great companies didn’t have better information than the comparison companies. The difference was “turning information into information that cannot be ignored.”[16] Visibility is irrelevant unless somebody is checking.

Measuring isn’t checking either. A client of mine had a rework problem that had grown to 32% over a couple of years. The shop did not fail to measure. They measured rework every week, more faithfully than most shops measure anything, and 100% inspection was already supposed to be happening. They were measuring it, but nobody was following up on the measure. Nobody was tracking why the rework was happening or acting on those whys. My advice had two parts. Get granular on the data, which defects, where, and why. And do just-in-time training, so when a problem is found, the person gets trained right then, while the job is still in front of them. Then I gave the leadership team the checking part, a short list of questions to ask every week:

  • How many bad parts reached the next operation this week?
  • What is rework tracking at now?

The first question is the one that tells you whether the parts got better or the log just got prettier, because bad parts that reach the next operation get counted downstream, by the people who receive them.

32%22%
Rework dropped from 32% to 22% in the first week.

The shop had the same people, the same machines, and the same measure it had carried for 2 years. What changed is that leadership started asking.

Trust but verify. Looking at the number is not checking. Asking why is checking. Larry Bossidy describes a CEO who chewed his executives out when they didn’t make their numbers but never asked why they didn’t make them.[9] That CEO was looking, not asking.

And give the checking its own check. A VP I know had operations getting marked complete incorrectly, which corrupted the data feeding the schedule. So his shop runs 3 levels of checking:

  1. A manager reviews every labor booking before an operation is marked complete.
  2. An automated check verifies the reviews are happening and sends an alert when they aren’t.
  3. He follows up personally when that alert fires.

A checking routine is a practice like any other, and an unchecked practice decays. You don’t need his automation to copy the idea. Put your questions in a fixed weekly slot, the same meeting or the same morning every week, so a skipped week is as obvious as a skipped review is in his shop.

Two stories show what this part looks like done well and done badly. Done well is a store in Irving, Texas that Sam Walton describes in Made in America. The store was a mess, and part of the problem was that no one was checking. “No one was checking on the refunds. No one was checking on the layaways. No one was even checking on the cash registers. If you wanted to steal, you knew you wouldn’t get caught.”[21] So they started checking on all those things, along with training, goal setting, and talking about integrity, and within a year and a half the store had turned around completely. Walton’s lesson from it goes back to his variety store days. “you’ve got to give folks responsibility, you’ve got to trust them, and then you’ve got to check on them.”

Done badly is a story from my own early consulting years. I had just landed a new client, a nonprofit that collects and resells donated goods, and another consultant was running a training session while I listened. Going around the room, he came to the woman responsible for loss prevention, who had been with the company for 17 years. He asked how much loss they had, and how much of it was theft. She said 30% to 35%. And he said, “What have you been doing for 17 years?” As if a number that ran for years were her fault alone. She was absolutely shocked and crushed, other people in the room were appalled, and the project ended right there. The check was aimed at the wrong level, because a number like that is allowed to continue for years only if the people above her never checked it, and that responsibility goes up, not down. When checking shows up years late and lands as blame, it becomes exactly the confrontation people dread. The checking in this system is small, frequent, and aimed at the system first.

Hardly anybody does this part. The authors of The 4 Disciplines of Execution report that 81% of the people they surveyed said they were not held accountable for regular progress on their organization’s goals.[10] So a shop that does check is doing what 81% of the people surveyed said wasn’t happening to them. The same authors quote business writer John Case on plants that put their metrics up on whiteboards. For a while, the numbers improved. Then a week went by without anyone updating a board, then a month, and the boards fell into disuse and came down. What gets measured does get done, but only for a while. A Velocity Board stays alive for one reason, and the reason is not the board. Somebody asks about it every day.

5. Improve the system

Even in a shop running all of this, something will slip. Find out why first. For scheduling, take the operator who keeps picking the easy jobs. Is that defiance, a sign the priorities need fixing, or not understanding what the board is asking?

And ask the why before you buy the fix. A client’s team wanted shop carts, because full kit kept failing and carts would organize the material. Nobody had asked why full kit was failing. The answer turned out to be that the tooling data the kits depended on had nobody keeping it current. It was nobody’s specific responsibility. Carts would have organized the mess beautifully, but they would not have touched a single thing that was causing it.

Here’s one from a client recently. A job needed rework, and the pipe cutter the rework needed broke. The team parked the job and ordered a new cutter, with a two-month wait. You can buy a pipe cutter locally. The only why anybody wrote down was “the tool broke.” Not one manager probed further, so the job just sat.

So what does VSS itself do with a slip like that? In VSS, the system can’t enforce itself, and it doesn’t have to, because VSS carries its own continuous improvement process, called POOGI, the Process of On-Going Improvement, named for the subtitle of Goldratt’s book The Goal. Every disruption to flow gets tracked, and the biggest disruption to flow is where the next improvement effort goes. I published that rule years ago. “We track the major disruptions to flow and use those to direct our continuous improvement”.[17] You want the same thing from Compliance Leadership. When checking finds a slip, the slip is not just a discipline problem. It is an input. Checking is not the price of the system. Checking is how the system gets better.

Then close the loop with the person. Whether what you found is a learning opportunity or a system change, circle back and tell them what happened to their idea or their workaround. If it was defiance, run can do, will do, fit. And if the slip traces to a manager who stopped checking, look at who manages that manager. Responsibility goes up.

6. Grow people

Who needs to understand the why? Managers do, to really drive results. For operators, understanding is optional if they follow the system. But the operators who want to understand are the keepers. For scheduling, the operator who asks why the board runs the way it does is a keeper. Watch for those people, because they can move into roles where improving the system is part of the job, like scheduling or engineering.

I learned early in my career what a judgment role actually does. At a charcoal manufacturing plant, they told me the process was all black art, that you had to know exactly how much oxygen to give it and how to cook it, and there was no binder that held any of it. What I captured there was not so much the tribal knowledge itself as what I could distill down to rules. Turning black art into rules the floor can follow is what a judgment role does, and the keepers are the people who can grow into doing it.

And judgment is about to become the scarce thing in your shop, because AI can output more information than people can absorb and judge. In my Critical Systemic Judgment article I define it this way. “Critical Systemic Judgment (CSJ) is the capability that determines which problems are worth solving, how they are framed, and whether a system’s design produces throughput or noise.” The keepers are your pipeline for it, and the daily asking is how you find them. I’ll have more to say about judgment in Part Two of my manifesto. While you grow the keepers, paint the vision for everybody, the end state of flow we’re trying to achieve, because the vision is part of the culture and part of passing down the causal model.

Start Here

Everything in this article works for whatever you’re trying to make stick. A scheduling system, a procedure, a quality initiative, a project. The split between results for the system and compliance for the people doesn’t change, and neither does the leader’s half. The 2 Laws say the same thing. “It is our job as managers to put systems and processes in place that get the results we want. Our systems and processes comprise the policies, procedures and measures that direct our peoples’ actions. If you’re not getting the results you want, just look in the mirror.” Don’t blame anyone but yourself for the results you get or don’t get.

And you don’t have to build all six parts this week, because the practice starts with one question. Here are the steps:

Start Here
  1. Pick one standing policy, procedure, or measure that is important to your system.
  2. Then start asking. Ask every day for a month and see what happens. It’s very likely that your priority will become their priority.
  3. Give it 30 days, then you can ask weekly but don’t ever stop asking.
  4. When that one holds, add the next one.

The system doesn’t change. Where you put your attention does.

Why only one at a time? Because you can’t check everything at once. Try, and you end up checking nothing well enough to change it. Stay on one policy until the new channel wears in, then take the next.

Does it work? There are two kinds of proof. The first is proof that the asking works. That rework client hadn’t even started the just-in-time training when rework dropped from 32% to 22%. All they had done was start paying attention to, and asking about, the 100% inspection that was already supposed to be happening.

The second is proof that VSS works, and the results are published. In a study of 442 job shops implementing the Velocity Scheduling System, shops saw:

198%
mean increase in productivity
42%
mean increase in on-time delivery
87%
mean reduction in WIP
82%
mean reduction in lead-time
Study of 442 job shops implementing VSS [22]

And more than 550 custom job shops have implemented VSS.

So start today by answering one question for yourself. In your shop, who is checking today? And if that person left next month, who would be checking then? If the answer is a role, you have a system. If the answer is a name, or nobody, you now know exactly what to build.

That’s what Compliance Leadership is for. If you have something you want to make stick, a procedure, a quality initiative, any project, we’ll put our logical thinking caps on and help you build out your compliance plan, like the client who cut rework from 32% to 22% in the first week, just by asking. Reach out to discuss it, and tell us what you’d like to build compliance for and where you’re struggling.

A great way to practice Compliance Leadership is with a project like VSS, so fill out a strategy session form if you’re ready to start or re-start Velocity Scheduling.

Frequently Asked Questions: Compliance Leadership

What is compliance leadership?

Compliance leadership is the method for getting your people to follow your system, your policies and procedures, and keep following them. The system is measured on results. The people are measured on compliance to the system. Managers design and implement the system, so the results belong to them, and if the system is followed and the results are still wrong, the problem is the system.

How do you hold employees accountable without being confrontational?

You check, and checking is two actions. You ask questions, and you go look. A question about the procedure, asked again and again, tells everybody the procedure matters. Holding people to the system is firm but not confrontational. You stay on top of it and keep checking until it becomes the new habit. It’s asking, not policing.

What should you do when employees aren’t following procedures?

Find out why first, and have the conversation. A workaround is typically one of three things: defiance, a sign the system needs fixing, or a lack of understanding of what the system is asking. Most turn out to be people who didn’t understand what the system was asking. A sign the system needs fixing becomes a system change, and a lack of understanding becomes a learning opportunity. For defiance, ask: can they do, will they do, and is there a fit?

Should employees be held accountable for outcomes?

It depends on the role, and the test is line of sight. An outcome belongs in a person’s role only if that person can touch or influence it. Upper-level managers can see and move the results, so they can and should be evaluated on outcomes. Floor roles are measured on compliance to the system, because what we do is counterintuitive and nobody on the floor would figure the method out on their own. It is wrong to evaluate people on an outcome they don’t know how to achieve and can’t directly affect.

Do bonuses get employees to follow the system?

Generally, no. Incentives become entitlements, and they break the employer-employee relationship: I pay you for your best work every day. A bonus pays for a number. A bonus plan belongs only with people senior enough to be responsible for the systems and the output, with the measure directly tied to that output. For hourly people, make them the best paid in the area, with the right measures in place and a good visual system.

References

  1. Eliyahu M. Goldratt, TOC Self Learning Program, session 7, Managing People.
  2. Eliyahu M. Goldratt, Necessary and Sufficient, discs 7 and 9.
  3. W. Edwards Deming, Out of the Crisis, 1982/1986.
  4. James Clear, Atomic Habits, 2018, ch. 12.
  5. Eliyahu M. Goldratt, foreword to a Japanese edition, c. 2008.
  6. Eliyahu M. Goldratt, The Haystack Syndrome, 1990.
  7. Eliyahu M. Goldratt, “Empowerment: Misalignments between responsibility and authority,” white paper, 1996.
  8. Elliott Jaques, Requisite Organization, revised 2nd ed., Cason Hall & Co., 2006.
  9. Larry Bossidy and Ram Charan, Execution: The Discipline of Getting Things Done, 2002.
  10. Chris McChesney, Sean Covey and Jim Huling, The 4 Disciplines of Execution, 2012.
  11. Kim Souza, “Store walks never go out of style for retired Walmart execs,” Talk Business & Politics, August 4, 2019.
  12. Robert Rodgers and John E. Hunter, “Impact of Management by Objectives on Organizational Productivity,” Journal of Applied Psychology 76(2), 1991.
  13. Eliyahu M. Goldratt, Isn’t It Obvious, revised ed., 2010.
  14. Efrat Goldratt-Ashlag, “The Layers of Resistance: The Buy-In Process According to TOC,” Theory of Constraints Handbook, ch. 20, 2010.
  15. Eliyahu M. Goldratt, “Comfort Zones,” 2007.
  16. Jim Collins, Good to Great, 2001.
  17. Lisa Lang, “The case AGAINST new Technology,” Velocity Scheduling System blog, 2010, https://www.velocityschedulingsystem.com/blog/the-case-against-technology-imts/
  18. Jeffrey Liker, The Toyota Way, 2nd ed., 2021.
  19. Donald T. Campbell, “Assessing the Impact of Planned Social Change,” 1976.
  20. Robert Reed, post on X, October 16, 2026, https://x.com/Oldhandai/status/2111195183140511799
  21. Sam Walton with John Huey, Made in America, 1992, pp. 122-123.
  22. Study of 442 Job Shops, November 2007 to April 2021, published results, https://www.velocityschedulingsystem.com/blog/velocity-scheduling-system-reviews/

Pin It on Pinterest