What Would Deming Say?

Introduction:

This week, Kelly Allan—a consultant who specializes in sharing the principles espoused by the late management guru W. Edwards Deming—returns to the podcast for a conversation with Paul Downs, Jay Goltz, and Laura Zander. After World War II, you may recall, Deming was sent to Japan, where he was largely credited with resuscitating the devastated economy. He of course went on to become tremendously influential here, too. And if you read his books or scan his “14 points” for management, it’s clear that many of his lessons are now widely accepted. But not all of them. For example, he encouraged business leaders not to set production quotas, not to hold people accountable—at least not without first holding the process accountable—and not to address employee performance and pay in the same conversation. Some of these issues came up in an episode that Paul, Jay, and Laura taped in December, which is why we decided to invite Kelly, who is chairman of the Advisory Council of the W. Edwards Deming Institute and has his own management consulting business, to join us. The goal was to see if we could figure out what Deming would tell Paul, Jay, and Laura, and whether the three owners would be open to his suggestions. Spoiler alert: Paul’s not really buying it.

— Loren Feldman

Guests:

Kelly Allan is founder of Kelly Allan Associates.

Paul Downs is founder and CEO CEO of Paul Downs Cabinetmakers.

Laura Zander is co-founder and CEO of Jimmy Beans Wool.

Jay Goltz is founder and CEO of The Goltz Group.

Producer:

Jess Thoubboron is founder of Blank Word Productions.

Full Episode Transcript:

Loren Feldman:
Welcome everyone. Jay, you said something that first triggered my thinking about Deming, and that is that the key turning point in building your business years ago was the decision to put up a board in your factory space, where you kept track of the work being done by your picture framers. Every hour, someone would call out to them to see where they stood, in terms of getting the day’s work done. And I found that intriguing, especially since it was so important to the development of your business.

Jay Goltz:
Let me give you a little more essence to that. You’ve got to remember, or maybe you don’t know at all: I’m not an assembly line. This is custom framing. Some people have big pictures. Some people have little pictures. It was simply a way of letting everybody know, yeah, if you fall asleep at your desk for three quarters of the day, we’re gonna know as the day goes on. Because every hour on the hour, we’re going to ask you, “What did you get done?” And all of a sudden, productivity went up, simply because nobody wants to scream out, “Loren, what do you got?” “Nothing!”

I didn’t say, “You better get three done in the last hour,” or, “You better get two done.” But at least there was some ongoing noticing of what was happening. And the major thing was that, if there was a problem, this is what would happen, before I did this. We’d get to the end of the day sometimes, and production was off by 50 percent. You go, “What went wrong?” “Oh, we got some bad frames.”

Well, it was too late to do anything about it. At least if it’s on the hour, if there was a real problem, you could immediately go and address the problem. And I’m not exaggerating when I tell you it changed my business. I haven’t had a quote-unquote bad day in 30 years. I used to have them whenever I’d hire a new manager. After about the fourth day, it was like a substitute teacher. They’d wait and see what they could get away with, and then production would slow to 50 percent. And it happened numerous times.

Loren Feldman:
Kelly, any thoughts?

Kelly Allan:
All of the folks we have here today are experienced business people who have obviously been doing so many things right. And with Deming, it’s not about what’s right or wrong. It’s about what’s useful, what has the greatest utility for return on time and investment, et cetera. But the concept of pay for performance is not really in this example, right?

Loren Feldman:
Right, we’re gonna get to that though.

Kelly Allan:
You didn’t really even set quotas, or say, “This is what you have to do to keep your job here,” which is another issue. What Deming is after—and I’m not channeling Deming, so it’s not what Deming would say or do. You’re hearing it from me based on what I know, so take it or leave it—but it’s the engagement with people. It’s working on their work together and to make sure that we’re really understanding the capability of the process.

And so whether it’s a known, documented process or not, there are steps. And every one of those frames, and every one of those things that they’re doing, that they learn pretty quickly and it becomes a part of their mastery of the skill set. So they’re probably working to their capability, and you’re not incentivizing them or berating them for not meeting a quota. You’re trying to understand what was causing that wide variation from 100 percent to 50 percent. That’s a huge, unacceptable variation. So you came up with what we call a common-cause solution—that everybody’s a part of the system, and they’re all doing the same solution. And unless you got some really negative feedback, it’s probably working okay. Does that make sense?

Jay Goltz:
Sure. Now, I read the 14 points of the Deming process, and I can’t argue with any of them, except one. I’m just scratching my head. And I don’t know that I understand the—it says, “Get rid of merit raises,” or something. What is the philosophy with giving raises? That’s what I’d like to know.

Kelly Allan:
There are basically two things about pay, first of all, just to set pay, which is: What’s the marketplace? Are you going to pay at, above, or below market? And then we don’t want the same job to have huge incremental differences in pay. So some people have been doing the job longer. They might be faster at it. They’re better teachers to other people. It’s hard to know. So what he’s talking about there, in part, is pay for performance.

So if I have people who I am paying merit pay because I believe their performance is better, I’ve opened myself up for a huge number of negative unintended consequences, because you only have to be wrong once. And if all the other workers know that that person’s cheating, cutting corners, cherry-picking the easiest jobs—I mean, there are dozens of ways to game the system. If everybody else but the manager/owner knows, you’re wrong once and you lose your credibility. And you increase the gaming of the system, because now everybody knows, to get ahead, that’s what has to happen.

Loren Feldman:
I believe in that podcast [episode] that I’m referring to, Laura, you talked about paying for performance in your factory in Dallas, where pay for the people who are making yarn there is based on how much they make. Can you describe how that works?

Laura Zander:
Sure, and our situation is a little bit like Jay’s in that we were trying to stabilize or provide some sort of clarity as to what was going on, because it was a business that we had taken over. So we were trying to get a measure of what our production was. And so what we did is, we realized that the first step in the production is the dyeing of yarn, and that there were multiple components to whether or not someone was productive, let’s say. And one component was how quickly they could do it: How many units per hour they could produce.

And then the other component was skill level. There were multiple different kinds of dyeing techniques, going from simple to complex. So we created a matrix where the highest-paid people are the people who can do the most complex work and produce the most at the same time. And then the entry-level is someone who, obviously, is slower and can only do the simple stuff. And our goal was to communicate to all of the people there that, “Yes, being highly-paid, and doing the most complex is the pinnacle. But you’re also really valuable if you can produce a tremendous amount at the simple level. And/or you’re tremendously valuable if you can do the most complex stuff, but you’re not super fast.”

And for our team there, we felt like it was important to put numbers to it, and say, “Okay, here’s the salary range.” Or, “Here’s the pay range for this job, and if you’re the slowest, most simple, this is how much you’re going to get paid. And then here’s kind of the matrix.” And we experienced gaming of the system, we experienced all kinds of stuff. And then we put some assumptions in place that all of this is based on an assumption of a certain error rate.

All of this is based on an assumption, obviously, that you’re a good cultural fit, and all these other kinds of things that need to go into it. But what has been interesting, is we put this into place about a year ago, maybe nine months ago. And part of the reason that we did this was to figure out as new owners walking into kind of a messed up situation and kind of a culture that didn’t fit with what we wanted it to be, we needed to figure out which dyers were doing a good job and which ones weren’t.

So we needed to put some metrics in place. As that has shaken out and as the pay has stabilized, we’ve ended up giving I think an average $3- to $4-an-hour pay raise, now that we know what people are capable of. One, their rates, the productivity, increased tremendously, as a result of creating some transparency. But then what’s happened now that we kind of have a stable benchmark and everybody knows—like with what Jay has—we’re shooting for 12 and a half skeins per hour, it is what it is. That’s the benchmark. If you’re lower than that, it doesn’t mean you’re going to get fired. But it might mean you get some more attention so that we can get you to where you need to go, to where we need to be, so we can do projections and all that kind of stuff.

What’s interesting, though, is as I was reading through those 14 rules, Deming’s rules, there was a lot of “replace with leadership, replace with leadership.” We have now gotten to the point where we have a good solid team, and we don’t have to focus on those numbers quite as much, because we know now that we have a good leader in place, we know that if everybody’s working really hard, they’re doing a great job, for the most part.

But we can still use these numbers as a way—especially when we bring somebody new in—to help them understand what the pace needs to be. I mean, I always liken that stuff—to me, I was a runner—I need to see my splits. I need to see what my times are to be able to know if I’m doing better or not. Because you can have a day where you feel like you’re working really hard, but you’re not being productive. And then you can have a day where you feel like you’re not on it, but you end up being more productive than you realized. So we think those metrics are really important.

Kelly Allan:
Yeah, so it’s interesting, because one of the things that Deming talks about is the understanding of the cultural piece and how to help people be successful. So with Jay’s situation and with yours, I think there are a number of things that you shared, which is: With the intention to try to make people successful with their work, with the types of jobs that both of you have, some people are just not going to be skilled at them. They just don’t have the dexterity, they don’t have—I don’t know what all is involved with it. They might not have the hand-eye coordination, the color sense. I mean, I don’t know. There are some criteria that would rule somebody out from being a good fit in that job.

And you mentioned some interesting things that are very Deming, which is—and Jay had the same thing—you can’t have variation of 50 percent and consider that you have a stable system. And you can’t really improve a system until you remove the things that are making it unstable, so that then you can continue to improve it in some different ways.

The only thing that might give me pause is, if you had it to do over again, one of the things that I would encourage you to look at is, “What is the capability of the processes in which people work?” And there are some ways to measure that, both with numbers and otherwise. I think you figured that out over time, without even realizing that what you were looking for was a metric for capability.

Because one of the things that helps us make more money faster, higher productivity, and greater joy in work comes from having above average processes so that everybody can get above average results. It opens up my hiring pool a lot more. Now, again, we’re talking about dexterity, etc., or some experiential pieces that may be a part of that. The other key piece there that you mentioned—and obviously you have a sense of this—is complexity. There’s a lot of research that shows that other workers respect the fact that someone who can handle greater complexity can get paid more money. So it’s not really merit pay.

I think in the transcript—Loren’s correct—the conversation was about merit pay. But what you’re really talking about, it seems to me, now that I understand it better, is you’re really talking about skill-set complexity and maybe some additional ability. When you get to an assembly line kind of operation or a service line kind of operation—and we can talk about some examples of that in a moment—then what you want to make sure of is that you’re not blaming or rewarding workers for the output of the process. Because that creates tomorrow’s problems today.

Paul Downs:
Is there a company that is well-known today that embodies the Deming principles?

Kelly Allan:
I love that, because it allows me to open up a quote from Deming. So there are hundreds, thousands—Toyota has gone a long way on the Deming journey. A number of other Japanese companies as well. And there are a number of U.S. organizations that use many of his teachings. But here’s what we’re discovering. So many organizations have started to adopt those strategic leadership pieces, even without knowing they originated from Deming.

Three of my nieces who are at various age groups will not work for organizations that do silly things that are command-and-control based. They won’t work for organizations where there are quotas that don’t make any sense, or where organizations cause them to compete against their colleagues. They’re working for organizations that understand that collaboration is what gets people to work for us, keeps people working for us, and gets better results, because people will share their ideas.

So when Deming was asked the question, “Dr. Deming, how many companies are using your methodologies 100 percent?” And he had a really deep voice, and he said, “None.” “Well, Dr. Deming, if that’s the case, in 100 years, how many companies will be using your methods?” “All that survive.” And that sounds like hubris and arrogance, but he ran the numbers, and he understood that if you’re at 3 percent today in the pond of algae in the pond, every time you get a double, pretty soon, it doesn’t take that long to get the whole pond.

We have started to see at the Deming Institute, and in my own company’s practice, we’ve started to see just huge numbers of things that, when Deming started talking about them in 1980, were considered to be crazy: “What do you mean, you don’t hold people accountable?” You hold the process accountable. “What do you mean, you don’t give quotas to people to make their numbers? What do you mean?” Well, that is all starting to seep in in big ways. Anything to do with technology companies. There are a few holdovers from that, but employees set their own vacation time, they set their own sick leave time, their own attend-a-funeral time, they do all of those things. And what the research shows is they actually don’t take as much time as the owners would like them to take, and the owners don’t want them to get burnt out.

Jay Goltz:
Can we get back to my basic question of: If you don’t pay people more because they are quicker or do a better job or have a better skill -et, how are you giving raises out?

Kelly Allan:
Well, I think in your situation, if it’s working for you, I wouldn’t tamper with it. In most situations, though, Jay, it doesn’t work very well. There’s constant sabotage of other workers. People don’t get along, which hurts productivity. And so it’s not about skill. It’s not about dealing with complexity.

Here’s a typical spreadsheet of what happens with most managers. So you have four workers, and here’s their production on day one. This is the actual number of errors, the number of defects they’ve made. And then their production of each worker on day two, day three, and day four. So what happens is, if I’m a supervisor, I end up looking at worker four on day two, who got seven, who was the best performance of the day—praising that person, rewarding that person, giving him a merit increase, and threatening worker number three, because they’ve got 15 defects. So take that to the next day, what happens is now the worker that I praised has now gotten 13, and the best worker now is worker number one—whether it’s daily, weekly, monthly or whatever.

Jay Goltz:
Are you talking about bonuses? When I hear merit, that means paying people more when they are, in the long term, doing a better job, versus what you’re talking about. To me, that’s bonuses.

Kelly Allan:
We typically don’t refer to that as fair. It is a level of skill, which requires a separate category of pay.

Jay Goltz:
Right. That’s not merit pay?

Kelly Allan:
Merit—typically, in most organizations—is used to praise, reward, or withhold merit for punishment.

Jay Goltz:
Okay, that’s very different.

Laura Zander:
So are you saying we should not have a meritocracy?

Kelly Allan:
Singapore, for example, for years has been a meritocracy. It works pretty well there. The entire system, though, is geared towards that. The piece that’s often missing is owners will grab— leaders, managers, owners—so-called best practices and try to cobble them together. And they can make the work. The question is: How much energy does it take? How much time does it take? Is that your highest and best use? Or is it simpler to simplify things, rather than creating more complexity?

So I would say to both of you, from what I’ve heard so far, both you and Jay are trying to simplify, trying to recognize the complexity and pay accordingly so that you don’t end up with a situation where the worker who was great one day stabs you in the back. That’s what managers feel. “They stabbed me in the back the next day, after I praised them.” So that’s the other thing that sort of popped out to me about both of your stories is, you’ve been trying to help people be successful, rather than being sheriffs. Now, if someone really cannot do the work, they’re unhappy, and you’re going to be unhappy. And so why prolong it?

Laura Zander:
We always say, “Our goal is to pay you more. Help us figure out how to pay you more.”

Kelly Allan:
Yeah, the only other thing I might suggest there is, try to get money off the table. Don’t always bring the focus back on money. What you’re trying to do is help them build skills. You’re trying to help them get more joy in work, because usually handling more complex work is more interesting after a certain point. People start to get bored. So let’s start to couch some of those discussions around joy in work and satisfaction in work, rather than just about the money.

Paul Downs:
Hey, I have some questions. In what world is it okay for someone to make 17 mistakes a day? And even if the best performance is five, what’s the total number of cycles completed per worker? Who designed the system? Who’s measuring it? Who’s in charge of making sure the data is correct?

I mean, we seem to have been dropped into the middle of a very complex way of thinking about the world that really only applies to certain kinds of work situations too. And I hate to be that guy, but this means nothing in my company.

Kelly Allan:
We always need that guy, first of all.

Paul Downs:
Well, you’ve got him. Get ready.

Laura Zander:
We just don’t always want him.

Kelly Allan:
So one of the things you said is: Who designed the system? Was it the workers? Typically not. So what happened? These are the kinds of things that we see. I mean, Paul, we see this day in and day out, in companies in every industry.

What we see is people will take the average, and that now becomes the quota. Where did they get their math training? It makes no sense. You have to take my word for it that this is very common. And so what happens is, there are only three ways to get better numbers. Only three: manipulate the numbers, happens a lot. Manipulate the system that gives you the numbers, so don’t report everything. Or improve. That’s it.

Laura Zander:
Why can’t the average be the average quota? I mean, if you’re finding on average that we’re doing 10, then over a year, can’t we assume that it’s going to be 10?

Kelly Allan:
Well, it depends on what your aim is. If your aim is to try to get the number of errors down to five or two, having a quota guarantees you will never accomplish that. Because you don’t know what the capability of that process is.

Deming was very careful with his words, but he said, “Quotas like that are a fortress against improvement, of ever really understanding what’s going on.” And in the meantime, you’re praising this person for low errors, punishing this person or even replacing them, but the capability of the system that they’re in, the tools that they have, the raw materials they’re working with, the process that’s been designed, the layout of the workspace, the lighting, dozens of inputs are causing that capability. That’s the voice of the process. The voice of the process tells you that any worker in that process is likely to pull between one and 17 errors a day. That’s the capability.

I’m probably the only one here who is old enough to remember Osborne Computer, the first laptop computer company. I mean, they had the market to themselves. They’re not around anymore, because they had quotas for everything, and many of them were averages. But at the end of their days, they were shipping empty boxes, because that’s when the salespeople got credit for the sale and when the production facility got credit for the shipment. They were shipping empty boxes and then getting on the phone and calling their customers and saying, “You’re gonna get 400 empty boxes, but I’m good for it.”

Jay Goltz:
I think part of the problem here is, the three of us are hands-on business owners and some of these examples are from huge corporations that are disconnected from a whole lot of stuff. And much of the stuff that you’re talking about is obvious to us, because we are the hands-on business owners. And this is as much about how to get to be a big screwed up company as it is about anything, and none of us are that big to get that screwed up.

And I’m in a weird business, in that, my defects? I’ve got a pissed-off customer. I’m not putting widgets out. My defect rate, quote-unquote, is next to nothing, because if I go ahead and something goes wrong, we’re standing there at the pickup counter with a customer who just spent $375, and they’re not going to be happy. So you talk about in-your-face quality control. By the nature of my business, we’ve had to make sure that we have people who are inspired and are dedicated to doing a good job, right off of your list. I mean, I do everything that’s on your list, because that’s the only thing that would work. I can’t yell at them to get it to be better. They have to care about what they do.

And so I can’t argue with anything that’s on your list. I just want to hear you say—because if I was the 26-year-old Jay going crazy with management and didn’t know what he was doing—I would want someone to actually say out loud, so we’re not confused, “Yes, sometimes, some people are not suited for the job. And your job is to figure that out and fire them.” I would just like to hear someone say that out loud.

Kelly Allan:
I thought I said that at least twice.

Jay Goltz:
No, you did. I’m just saying, when I used to go to seminars, I used to think, “What am I doing wrong?” Because no one would tell you the truth, and I just wanted the truth to be out there.

Kelly Allan:
Deming said the same thing.

Jay Goltz:
Good.

Kelly Allan:
In fact, there’s pretty much a rule that says, “We will try to help you multiple times to make sure that we’re not blaming you for a process issue. But if we’re pretty sure it’s you, and if we’ve tried several times, then you’re not a good fit. You have to go away.”

Jay Goltz:
Wow, then I am listening to all of the principles, because I say after the third conversation that I think three times is probably enough to figure out that they’re not getting it. So okay, good. I’m all in.

Loren Feldman:
The clock is ticking and our hour is passing. There are a couple of things I want to hit. I want to address something that applies directly to Paul, and I think this may get to some of the questions that he started to ask before. Paul, in that podcast episode that I was referring to earlier, you talked about the problems you had in your custom manufacturing process—because there’s a wide range of possibilities. There are a lot of external factors that could have an impact on how long it takes you to do something, to build something. It affects how you price what you do, it affects your ability to figure out what your margins are for any given piece that you’re making. Paul, do I have this right? Could you talk about why you struggle a little bit with external factors and not knowing for sure how each project will go?

Paul Downs:
Sure, in case anybody doesn’t know what we do, we build custom conference tables. There’s huge variation in what we’re actually doing on the shop floor, and generally, we also have somewhere between 20 and 50 jobs on our shop floor at any given time.

And so it’s a lot of different stuff, and a lot of it happening at once. And it’s pretty hard to apply these concepts that were developed for assembly lines and put them into a more chaotic environment. And I think that my dilemma with these systems is they play out in a lot of different businesses, that what people are doing all day may not be easy to measure. The defects may not be easy to measure, and also the make-or-break activity that makes the business succeed or fail may not be what you’re measuring.

So just to go to Jay’s example, he’s measuring people putting together picture frames, and it may well be that the make-or-break activity of his business is whether the door got unlocked on time—the front door—or whether the ad has a misspelling in it, or whether the person who picks up the phone is a misanthrope, or any number of things that are happening way off in the distance, and may or may not have an effect on what’s happening wherever you choose to focus your attention.

In my business, we try to measure what’s happening on the shop floor. And the closer you look, the harder it is to make any sense of it. But we do collect a lot of data so that we can come up with aggregate numbers. And what I was saying on the podcast episode was that I have a couple of numbers that are just aggregations of performance, really of the whole company.

So I talk to my people about revenue goals per month, because it costs more or less $350,000 a month just to run the factory. So if you have revenue of $380,000? Happiness. If you have revenue of $330,000? Bad. That’s really simple, and so I talk to the people about, “Here are the jobs that are coming up, here’s what they’re worth, here’s where they need to go out, you figure it out. I’m giving you information. This is where the finish line is. This is how we succeed. This is how we fail.” And then all of those sub discussions about how our processes are happen in the context of what the overall goal is.

And second of all, what we’ve talked about so far seems very applicable to an assembly line, and not necessarily to any other situation. So is that an answer to your question, Loren?

Loren Feldman:
There’s a lot there. Kelly, would you have a reaction to any of that?

Kelly Allan:
I need to clarify. I mean, it works on assembly lines, but most of the organizations that are using this are not assembly line work. They’re much more custom kind of operations, many of them service, certainly some manufacturing. Manufacturers picked up on Deming a lot at the beginning. And Deming would certainly applaud, I believe, looking for simplicity. And that’s what some of these insights help with, is how to make it more simple. And he also said—I don’t know if you’ll appreciate this or not, Paul—”The most important things that we need to know as business owners are unknown and unknowable. They’re just too hard to measure or impossible to measure.”

So when that’s the case, what we need to do is, we have to look at: What is the theory? And so part of that theory is: What is the least we can do? What is the most simple thing we can do and not spend a huge amount of time on trying to do spreadsheet analysis? Instead, take a few seconds, throw it on a chart, which will tell you the type of variation you have.

There are only two types of variation. And the one type of variation is to work on the process. The solution is to work on improving the inputs to the process. And the other type of variation says: What caused that? It’s worth the money to go investigate what might have caused that bad thing or good thing to happen.

But what happens to most managers, is they are taught that every data point is a special cause that they have to go investigate and reward or punish. That’s where the huge losses that we see in most organizations come from, of size, and we work mostly with small businesses. Yes, we do work with Toyota and other big names. But most of our business comes from small businesses. Helpful or not?

Laura Zander:
Paul and Jay, it sounds to me—when I read those 14, all I came away with was, “Treat your people right, and treat them like people.” And if you’ve got the right people, and they’re working really hard—and everybody’s capability is different—if you just spend time working on improving the process, as opposed to trying to make people who don’t fit fit, then the rest will just take care of itself. Sure, you need to measure some stuff, because you need to figure some stuff out and be able to project and do all that. But you’re not measuring to reward or punish. You’re measuring to help improve the process.

Kelly Allan:
That’s a key point.

Laura Zander:
The 14 just seemed like… I mean, I’m sure Einstein was really good at simplifying things, but it was just: Be nice to people, be good to people, and trust.

Jay Goltz:
I think respect. Let’s be respectful.

Laura Zander:
Yes, trust and respect. And that’s where the good leader comes in. If you have good leaders, then they will trust, and they’ll help people rise to their capabilities. But I don’t know, it seemed pretty common sense.

Kelly Allan:
Something that might be useful for the growth of your business, Laura, has to do with what you said, which is: What is the definition, what’s the criteria, for a good leader? What does good leader mean for your organization? Do they use command and control? Or are they encouraging and being sheriff? Or are they encouraging collaboration? Are they mentoring? Are they helping people get the tools that they need, the understanding of the job so they can be successful? Those kinds of criteria help save time and money when you’re hiring a leader or trying to mentor a leader, a manager.

Laura Zander:
Isn’t command and control like super 1950? I mean, do people still do that?

Kelly Allan:
Oh my gosh.

Jay Goltz:
I’m gonna tell you the question I always ask. I’ve done seminars to small business owners, and I say, “Let me ask you a question: Do you ever yell at your employees?” And I go, “Let me read your mind. You’re thinking, ‘Well, I do that. But that’s because I’m passionate.’” And I go, “No, it’s because you’re an asshole.”

Laura Zander:
Right.

Jay Goltz:
“It doesn’t pay, and stop doing it.” And I would say from what I’ve seen in businesses, the screwed up businesses, there’s a screaming boss there somewhere. And they never figure out, “This isn’t helping. It’s not working.”

Kelly Allan:
No, you’re absolutely right. We just participated in a big study, a university study of middle managers. And you know what they spent most of their time trying to do? I mean, it’s stunning, right? It won’t surprise you. But it’s just evidence. The top three things have to do with motivating people. And what you all are saying is what Deming said, “Stop demotivating them!”

Jay Goltz:
That’s exactly it. They always ask, “What do you do to motivate?” I go, “Well, there’s probably five ways to motivate. There’s about 100 ways of demotivating.” I said, “Stop doing that.” So, wow, it sounds like I’m a natural Deming kind of guy, then.

Kelly Allan:
You may be, but there will be no reward for that.

Loren Feldman:
Can we go back to Paul for a second? I’d like to get back to that issue of not being able to predict what is happening with specific projects, Paul, and the external factors that cause that. You brought it up as if it was a constant, ongoing thing. And I’m wondering how big a problem for you that really is. I mean, do you have instances where you take jobs that you regret taking because it turns out they cost more to produce than you thought they would?

Paul Downs:
Yes. Okay, so just to explain, we’re building a custom thing and one of the biggest issues in a custom world is just figuring out what the price should be. Because if you just say, “I’ll do anything,” chances are good you’re going to be doing things that you never did before. And it’s very difficult to figure out how long they’ll take or what the issues might be.

When we’re selling tables, we have a pricing spreadsheet that allows the salespeople to kind of build a table using features and algorithms that kick out numbers, and I wrote all that. And there are hundreds of different calculations in it. And how did I preload the answers in? What did I do to make sure the algorithms were right? Nothing. I just made shit up. And the reason is, it was actually impossible to run tests on all of the things that I needed to have calculations for.

So what I did was, we wrote the thing, and then we ran it for a while. And we looked to see what happened. What is the variation performance? How well does it predict? It doesn’t really do a great job of predicting any particular thing, but it does a good job of predicting the aggregation of all the jobs. And what’s happening in real life is, in our shop floor, we have a complex interaction between the thing being made and how complex it is. The material we use, which is wood, is highly variable when it comes in the door. One walnut tree is not the same as the next walnut tree. We have a range of people, and people come in a variety of skill-sets. And then on a given day, they may have a headache, or they may feel great, or they may have 20 cups of coffee, or they may… who knows? There’s not constant performance on them.

And then we have the clients and then the rest of the company. Each person is operating within a matrix of all of the other people and what they’re up to. So it’s a highly complex fluid situation. And what I found was, once we got an algorithm that was reasonably good, it was best to just throw it out there and say, “Here’s the expectation.” And we would see what the variation was, and we could easily identify by measuring when the algorithm and the reality were highly divergent. And we tended to concentrate on jobs where the algorithm had kicked out too few hours to complete the job. And whenever the algorithm kicked out way too many hours to complete the job, and the client ended up paying for a lot of unused time, we just call that a victory.

Even though, when you think about it, those are two sides of the same coin, which is the algorithm isn’t conforming to reality. And basically, my business survives whenever we can predict wrong in the right direction. And I don’t know. I mean, that’s just my life. And I’ve taken enough measurements over enough years to realize that, and we’ve actually been able to identify situations where the algorithm is likely to be wrong in our favor, but we don’t do anything to correct it. Because we know that’s where we make our money.

Kelly Allan:
I don’t know why you don’t raise the algorithm to do that all the time.

Paul Downs:
Because one of the things that’s very difficult is we’re selling a thing that nobody knows the price of, so that every job we do is a one-on-one negotiation with somebody who’s got some amount of money in their pocket. And we’re working against their perception of what this thing might cost, should cost, what else they’re looking at, in terms of, “Okay, if I don’t buy from Downs, what else could I do?” And that’s usually something that is crappy, but you know, it’s cheap. So we’re aware that they have other options, and we kind of have to make a decision about pricing at the point we’ve got to price it in order to get the job done.

We can’t be tied to just the algorithm. And yeah, there are definitely times if the shop is running out of work—and I don’t want to lay everybody off—I might buy work knowing that it’s not going to go well. But at least we don’t have to send everybody home. And there are other times when we know that the client is feeling happy and ready to spend, and we may go for every penny we can get out of that person. Because if they think it’s worth it, it’s worth it to them. So what can you say? Price is fluid. And so that’s our situation. So it’s not necessarily what happens at the shop floor that makes the job. It’s actually what happens in the negotiation between the salesperson and the client.

Kelly Allan:
That’s a nice way to tease out the critical piece there for me, which Paul was just talking about. There are some things that help with making and doing that are kind of more within our control. But there are those other things that will determine success that we may not have much control over. So the prospect’s budget, prospect’s perception of quality, prospect’s desire for a particular look, whatever it happens to be. And so then the business owner gets to and needs to make the decision you just said, which is the prediction’s good enough. In other words, you are making predictions—management is prediction. You are making predictions within the ecosystem, if you will, that you have designed, right?

Paul Downs:
Yeah.

Kelly Allan:
I don’t know if this is helpful or not, but all businesses, at some level, are generic. Every business takes some input, adds value in the throughput, and gets an output. I mean, it’s not really a lot more complicated than that at that 30,000-foot level. Where it starts to get more complicated is: How can we improve those inputs? I suspect you’ve been doing it a long time. You guys are probably very good at the negotiation. But if you haven’t looked at negotiation for a while, that might be upstream where more money could be made. I don’t know.

Paul Downs:
Well, thanks for that insight. [Laughter] Yeah, obviously, if we could get more money out of all of our clients, everybody’d be great. I think that’s true of all businesses, right? So easy.

Kelly Allan:
Well, no, that’s what I’m suggesting in terms of negotiation skill, right?

Paul Downs:
We’re pretty good at that.

Kelly Allan:
No, I suspect you are. You’re very successful, so I suspect you are. So what changed for you from the early days until now?

Paul Downs:
I would say that the biggest change was A) in my own capacity as a leader, but that took two forms: One was getting out of my own head and starting to make connections with other business owners, so I learned something, as opposed to just stewing in my own juices. The second one was that I’ve come to realize, or believe, that what you want is to make sure everybody knows the rules of the game they’re playing.

And this dovetails nicely with the 14 points, but we have expectations about what people need to be doing all day, and how they need to deal with their co-workers, and how they deal with customers, and how they deal with their work. And they’re very explicitly written down, and people are coached to them. And a lot of those things are just stuff that I believe, but those are the rules. And I think that most human beings are actually quite comfortable in a situation where there are very explicit rules, as long as they’re applied fairly and uniformly.

Then the other thing is that I’m very much about repetition of message, sort of like building rituals into day-to-day operations, based on the observation that the most successful human organizations are religions. And they do it by telling you the same stuff on a regular schedule. And so that I’m trying to tap into the basic human desire to feel like they know what’s going on. And then you can let people operate with minute-by-minute autonomy to get work done, as long as they’re working within an overall framework that makes sense to everybody.

And if people cannot operate within our rules, we get rid of them. And when I hire people, I say, “Here are the rules. Do this, you’re good. Don’t do this, you’re bad.” And we give people regular updates as to how they’re progressing to make sure that they understand what’s happening. That’s my system.

Loren Feldman:
In that podcast episode, Laura, you talked about how you track your financials and the amount of time you spend just going through everything. And you do it in part because you enjoy doing it. And Kelly, I think you listened to that or read that and reacted to it that it sounded like an awful lot of time, and perhaps not the best use of time for somebody running a business. Do I have that right?

Kelly Allan:
Yeah, my sense was of empathy.

Loren Feldman:
Keep in mind, she likes it. She enjoys doing it.

Kelly Allan:
I somehow missed that part. I don’t know. You get to decide. You would have better insight about what your highest and best use is. I think it was part of a discussion about whether you needed a CFO or wanted to look at a CFO or something like that. And my concern there was that, unless there’s a new method that the CFO is going to be bringing to the data to analyze the data that’s faster, better, cheaper, smarter, that it wouldn’t relieve any of your anxiety and not save you any hours. In fact, it might have the opposite effect. But I didn’t know.

One of my questions I think was, “Well, wouldn’t I want to be looking at sales all the time?” I can’t remember quite what the context was, in terms of the cycle at which you were looking at certain things. But I said, “If she had the time, she might be able to look at that more often. And that might provide greater ability to predict and less anxiety, less concern about what might be happening.”

Laura Zander:
Yeah, that’s interesting. I do look at my end-of-month numbers line-by-line, category-by-category in the P&L and in the balance sheet, every month. But it’s almost like a meditative experience.

Kelly Allan:
Interesting.

Laura Zander:
So I look at it from a place of curiosity. I’m like, “Oh, interesting. Oh, this is interesting. Oh, we’re using Canva now. Okay, I didn’t know that. That’s interesting.”

Kelly Allan:
You’re exploring.

Laura Zander:
I’m exploring. Yeah, I’m going through as an explorer, and then I’m looking. I think my brain is suited for seeing patterns. And so I’m looking at it from a pattern standpoint, and I’m looking at shipping expenses, [which] is the easiest way. I’m like, “Wow, shipping has continued to go up and go up and go up.” Again, yes, it’s exploratory. It’s very, very exploratory.

Kelly Allan:
For the patterns, are you using a visual display? Or is it spreadsheets?

Laura Zander:
Both, so sometimes spreadsheets. And then, now I have new software that’s allowing me to see some visual displays. I mean, I would love to see something like what you’ve got. Part of my issue—and one of the reasons I wondered about the CFO—is that I don’t always know what I’m looking for, and I don’t know what questions to ask. I’m just not experienced enough.

Kelly Allan:
I don’t want to discourage you from exploring that with a CFO. If you’re looking just to do an experiment with something that’s Deming-based with the visual display of data and how it can show you more at a glance for those line item budgets, etc., and what the guardrails are for that and give you a signal. There are other signals that appear to show you that this is headed out of line. This is trending in a direction. So the patterns are very clear on the dot plots.

Laura Zander:
Well, maybe, Loren, we could get Kelly to do like a pro bono analysis of our business, and then we could present that to the 21 Hats Podcast audience. [Laughter]

Loren Feldman:
I don’t see why not.

Laura Zander:
And, you know, do the whole journey.

Jay Goltz:
Makes sense to me.

Laura Zander:
I think so. Paul, what do you think?

Paul Downs:
I’ll go last.

Loren Feldman:
Much to my disappointment, it appears as though I’ve failed to instigate any real arguments or disagreements here.

Jay Goltz:
Well, I have an argument. I think “merit pay” is a bad phrase. I think what he’s giving is bonuses.

Loren Feldman:
I knew I could count on Jay.

Paul Downs:
Wait, before we argue about this, we should at least—let’s see. This is bup, bup, bup, bup bup: “Remove barriers that rob people in management and in engineering of their right to pride of workmanship. This means, inter alia, abolishment of the annual or merit rating and of management by objective.” Okay, that’s the actual wording of what you objected to, right?

Jay Goltz:
Yeah.

Paul Downs:
Yeah, I think that when you say, “Don’t have merit pay,” you’re kind of teasing that out of some words that may not actually be there. This strikes me as a commentary on sort of the General Electric management practice of firing the bottom 10 percent of the year. And that kind of, “Oh, we’re gonna give you A’s and B’s.” And the danger in those systems is: Who’s writing the report card? Who’s grading it? Is a bottle of whiskey at the end of the day going to give you an A? All of that corruption that we saw in the 1950’s auto factories, that strikes me as being what this is talking about, and not necessarily a broad statement that no one should ever get merit pay. Am I correct in that?

Kelly Allan:
Well, we’re back to the operational definition. If I’m a part of a system with other workers, and the manager is looking at comparing and contrasting two numbers, versus finding out if we’re all in alignment in the same system, that’s a problem.

Paul Downs:
And again, that goes back to my desire to make sure everybody understands the rules of the game they’re playing. And one of the rules in my company is: You don’t win or lose individually. You can be fired. But a mistake doesn’t necessarily mean that we’re going to come down on you. The only measure of success is the company. And if the company succeeds, I will be able to hand out all those goodies that people need. And that’s why we’ve simplified it to just a monthly revenue target, and then constantly tell people where we are in meeting that target. And I assume that if they can manage not to murder each other when they see each other every day, that they’ll figure out a way to work together and solve the problems that need to be solved.

Kelly Allan:
You mentioned something that most people don’t understand. Deming, he used the words, “What we want to make sure we’re doing is optimizing the overall organization.” And to do that, we sometimes have to sub-optimize certain pieces and parts of that. So it sounds to me as if, Paul, you’ve grasped that.

Loren Feldman:
Explain what you mean by that, Kelly.

Kelly Allan:
So we’ve typically been taught that if you optimize every department, you will optimize the overall organization, and I think these people will tell you that that’s just not true. That’s just not true. It’s not how it works. The way it works is, if we collaborate and work together with the flow of everybody contributing to the greater parts, that is the key. That makes all the difference.

A simple example that I think most people would understand is, there was a fad going around for a while in which every department was a profit center. This lasted seven-eight years. Every department had to be a profit center. I don’t care if you were janitorial, or the internal travel agency, or whatever you were. Because people are not going to collaborate. They’re going to become selfish, what’s called local optima. They’re gonna become very selfish for their own department because they’ve got to show their profit. They’ve got to show their merit. They’ve got to show how they’re contributing, rather than truly collaborating across divisions. Does that make sense?

Jay Goltz:
The way I summarize that whole thing is, I think that businesses use sports analogies too much. In sports, there’s always a winner and a loser. I believe I’m running an orchestra, and I want all the orchestra to work together, to play perfect music that sounds lovely. That’s what I want to do.

Kelly Allan:
That’s another Deming example. He said, “If you want a great orchestra, you don’t want everybody coming in, playing loudly and all the time.”

Loren Feldman:
Thank you, Kelly Allan, Paul Downs, Laura Zander, and Jay Goltz. I appreciate your allowing me to experiment with this. Thank you very much.