Transcript
FACILITATOR: HI guys. So we're here again and now we're here with Steven Ouellette and he's a consultant, educator and instructor with an engineering background. He has experience working with executives to turn their long-term visions into metrics and deploy their metrics all through the organization.
Go ahead tell us a little bit about us about you and take it away!
Steve: Hey welcome everybody! It's good to have you here. I’ll go ahead and be talking about “One Weird Trick” to help you create the right metrics “Doctors Hate It!”
Okay, so it's like the internet meme but it but it's an actual real thing. So let me introduce myself first.
My name is Steven Ouellette. My undergrad degree is in metallurgical and materials science engineering, which is a fascinating topic During the time that I was working as a metallurgical engineer at ALCOA, I met a consulting group that was doing the type of thing that we're going to be talking about today, and I realized as an engineer I knew a lot about the process but I didn't know a lot about the business. That got me really interested, so I took everything that those guys could teach me and that was back in 1996.
Since then I’ve earned my CMC, my certified management consultant marque, and I’ve got about 25 years of experience consulting. In 2002 I started my company the ROI Alliance. I’ve been a professor at the University of Colorado Boulder on both the Leeds School of Business as well as the College of Engineering and Applied Science in their Engineering Master's degree. I am the author of a couple of books: Business Performance Excellence was the first one, and just published in December, Galileo's Telescope. More on that in the next slide.
Some interesting facts about me: I’ve got a black belt in a Japanese martial art called aikido which is a non-violent martial art. I spent a year with the Thomas J Watson fellowship traveling through most of Western Europe studying the evolution fabrication and social impact of European swords and I spent some time in Spain learning about swords and sword making as well as most of the rest of Western Europe.
So that's a little bit about me. I come to this as an engineer. I come to this as a person who looks for logic in in how to do things, and there is a logic in running business and that's what we're going to be talking about today.
I just mentioned that I published a book in this December it's called Galileo’s Telescope. it uses the life and accomplishments of Galileo to kind of illustrate how you can use metrics in your business how you create the metrics and how you can then use those to lead and to manage a business. So what I’d like to do is if you post a question or even a comment in the chat you'll be added to a drawing for two free signed copies and we'll get a copy of that to you later.
So why is creating good metrics so hard? It seems like a lot of businesses and other organizations really struggle with this. With how to try to figure out what it is that is important to measure for the overall business and for each individual. It ends up being something that's almost overwhelmingly difficult, so let's take a look at what companies often do today and we'll kind of transpose that against how I would approach the metrics creation for organization.
Very commonly today, we've got no metrics, and I’m going to steal and modify a quote from Dr. W Edwards Deming saying if you have no metrics in your business, if you have no measures of success it is kind of like driving a car at night with the lights off by watching just the rearview mirror. It's not a very safe way to run an organization. You don't see what's ahead of you, you can't react in time, all you can see is basically where we ended up for the quarter or for the year. So this is common, but clearly not the safest way to run an organization.
Sometimes you'll see a top-down approach. Now there are dangers associated with this approach that I’m going to go through but the idea is the boss says, “These are the metrics that you need to measure, there you go just measure them.” There are problems with that, however.
The first is there's a lot of fear in organizations. People fear the use of numbers to beat them up or punish them for things that are outside their control, and they have a long history of this. Typically the people who are in our working in our businesses have seen either themselves or others be punished with a stick, beaten with these metrics in terms of performance. It's really important when you do put metrics together that they are not any part of a performance review. Performance reviews shouldn't happen anyway they're terrible, terrible idea, but for sure any process metrics that you put together should be process metrics. They don't reflect positively or negatively necessarily on an individual. We need to get that fear out of the organization so that we can understand what's going on in the organization and react dynamically to it as managers. We want to drive that fear out of the organization.
Also going top down leads to the risk of a danger of no translation. If the boss tells me to measure something, I don't know why it's important, I don't know what role this plays and what the organization is trying to accomplish. Your boss's metrics don't mean anything to you, and so I need there to be some sort of translation so that I understand why this is important to measure. But instead, I’ve just been given this with no explanation and so I don't understand why it's important. And I’m probably not going to be all that thrilled about doing it.
Another danger is that of non-participation. If a person isn't involved in the creation of the metrics for their own job then why would they follow that? If somebody comes to you and says, “Here, I’ve just met you for the first day. I’ve looked at your process and here are the things you need to measure with this process.” You probably wouldn't be very happy in hearing that, and so you would likely resist the implementation of these metrics because you have no role in their creation, no agency or ability to create or modify those metrics. And if you think about it the people who are who are actually running the process themselves are the process experts and oftentimes imposing metrics on them are going to be things that are not nearly as good as they could have come up with themselves. So there's a lot of resistance to these metrics being imposed from the outside.
Another one that I see a lot is the danger of constantly changing metrics. When CEOs read about a new metric in an article where they hear that Microsoft or Boeing or somebody uses this particular metric, then they try to push that metric into their organization where it may not make sense. And then they read another article next month, and then it's a different metric and then they start chasing after that one. There's no constancy in a situation like that - you're always chasing these new metrics and forgetting about the old ones. There are no stable goals - it's always different. It's a very confusing environment to be in.
Another risk with top-down metrics is the danger of the boss just focusing on one metric. Here we have got a picture of boss who thinks things are going great because he's only watching the one metric, but you can see that the business is falling apart behind him. He doesn't have visibility on the other important aspects of the business that he's there to run. If you only focus on one metric you can make really dumb business decisions. For example, if a company only focuses on profit, what about safety what about innovation, about environmental metrics? All these different metrics are not looked at because we're only focused on the one metric. I guarantee you if the boss is focusing on one metric, they're going to get that number, but the problem is at what cost are we doing it? It’s at the cost of the long-term viability of this company and its people who work there.
Now, new recently is another way of generating metrics, which is from the bottom up. So nowadays, there's a lot of data being generated in the processes themselves, and these data often go into big data warehouses or data lakes and then we can roll those metrics on up to the boss all the way up to the CEO.
Let's take a look at an example of that. This is an example of front-line metrics - people who are doing the process themselves - just being rolled up to the CEO’s level. As you look at this, you see that there's a lot of things that the CEO should not be involved with. If the CEO sees that the average satisfaction dips down to 92 percent he or she may feel obligated to intervene in the process, to change something, but not understanding if that's part of the normal variation or if somebody's already working on it.
So if the CEO has a metric like this and they react to it, they're jumping multiple levels of managers in order to talk to the people responsible for whatever this metric is. Which means all those levels, all those different levels of managers have just been disempowered. They've had their management objectives taken away from them by their CEO. We don't want that - that's called “micromanagement” and micromanagement is enabled by these abilities to take these data and roll them up into something for a CEO to see.
Now you can't blame the CEO: as they're watching these numbers go up and down they feel like they need to do something, particularly if these are going to turn red if they go below a number. The CEO feels like they need to make some sort of intervention. The problem is that in doing so, they're being consumed every day by a huge number of metrics. They don't have the ability or the bandwidth to provide the value that really genuinely they're there to provide. Their value is a strategic value, right? They're supposed to be looking out further in time across new potential markets, or asking “What is my business going to need a year two three five years from now that I need to start getting today?” If all they're dealing with is today's numbers that are going up and down, not only are they disempowering every other manager underneath them, but they're not spending time on where they have the most value which is in where we need to be going, not what's happening today.
So as you see things like business intelligence software it could be Microsoft's Power BI it could be Tableau, any of those things exist. That's good information, but you don't want that information to go up too high. Typically when I build some sort of a business intelligence display, I disallow any manager from going down more than one level for exactly this reason. I don't want them consumed with the tactical day-to-day data. I want them looking at the longer-term data and spend the time that they aren't doing that thinking strategically.
Now another way instead of from the top down or bottom up that I’ve seen is every manager for themselves, and this is probably the most common one in businesses that I’ve seen. Each manager just measures whatever they want about the processes in the area that they manage, and that's what they do. There's no communication across or up and down into the organization about that. So what would be some of the dangers associated with that?
There's the danger of an of an incomplete set of metrics. Here we've got an example of a bison, and these are the metrics that I’m looking at and things look pretty good where I’m standing right now. The problem is there's a bigger picture, and this tells you a completely different picture. If you get a complete set of metrics, you can understand everything about what's going on, not just a few things that you happen to be tracking, so incomplete measures may look good, but complete measures give you a better understanding of the real situation that you're in.
Another danger associated with managers creating metrics on their own is the danger of wrong metrics. Wrong metrics drive wrong action. If I’m measuring something, people are going to think that it's important whether it is or not, whether it's important for what the company is trying to achieve or not, and they will work to get you those metrics.
I think of an example from the former Soviet Union where aluminum plants had a target number above and below which they could not go for how much metal they were supposed to produce each month. If they produced a little extra amount of metal they'd dig a hole in the ground and put the aluminum into it and then cover the hole back over and get exactly the number. Then the next month when they're a little short, they go out there they dig up the aluminum, put it on the truck, and then they'd get exactly the amount. So the metric was driving real behavior, but it wasn't behavior that was adding any value to the creation of aluminum or the shipping or the alloying of it. It was only there to hit a number, and a lot of times you'll see these wrong metrics driving people to get numbers for the sake of numbers, when they aren't really very useful numbers, not really aligned with what the business needs to accomplish.
Another example of a danger is misaligned metrics. So have you ever seen one part of a company or one area of the company being given multiple awards for hitting their number when the rest of the business is kind of on fire behind them? This is showing that the metrics in one area are probably unrelated to business success. Sometimes it's even in the opposite direction.
Oftentimes individual areas should actually sub-optimize their performance. They should take a hit in maybe a cost or a production rate, something like that, because that's what the rest of the company needs. The classic example of this if you're in a manufacturing type environment is with maintenance. It's really easy to “save money” with maintenance by reducing the number of workers that I have, or by taking them off of shift work or by getting cheaper vendors for my for my parts. As I do that, it looks like I’m saving money. But of course, the consequence to the company as a whole means there's a lot more unplanned downtime, machines that are breaking when we didn't expect them to, which is very, very, very costly.
So where one area looks like they're doing great they're saving a lot of money it's actually costing the overall organization lots of money time and effort and potentially quality as well. We call this a situation of sub-optimization: optimization of an individual area leads to sub-optimization for the company as a whole. If you have misaligned metrics, if you have metrics that are working in opposition to other parts of the organization or to the organization as a whole, then this is what's happening. You're hitting your numbers but everybody else is unable to.
So those are the dangers, those ways that people currently try to create metrics and find them very difficult. How do we go about making good metrics? We'll spend the remaining time on that.
Making good metrics is a craft, and because I studied sword-making I had to get a picture of a sword for you. It's a craft, it's something that you work at, it's something that you can continue to work at and improve over time. Don't think that the first time you put together a set of metrics it's going to be perfect. You're going to have to learn from those first metrics and improve in the future.
Let me show you the process that I use, and which is outlined in my new book Galileo's Telescope about how I use metrics and create the metrics within the context of a business.
First off, let's talk about characteristics of good metrics. You need a good metric because metrics drive behavior. Wrong metrics drive wrong behavior, right metrics drive right behavior. So as they drive the right behavior, we need to use them in order to support making the right decision day to day every day, which is the role of management.
Good metrics have to be aligned. They have to look into the business requirements, into the business objectives, into the value that that business creates and be aligned with those things. If it's a metric that doesn't align with anything that the business cares about, why would we measure it?
These metrics need to be actionable and by that I mean the people who are responsible for the metric have to be able to take some action to respond to it if it goes into a bad direction or a good direction. They have to be able to do something to affect it. If you're giving people metrics that they can't affect, it's not a very fair thing. You're putting them in a situation where they're responsible for something, but they can't do anything about that thing.
The metrics need to be complete, and by that I mean they need to give the entire picture not just the one picture of the bison but the entire picture of what is necessary in order to run this particular area. If I’m missing a component, I’ll make decisions and it makes some metric go bad, but if I’m not measuring it, I don't know it that the consequences can be very bad.
I need to have a complete set of metrics that are all aligned and actionable. Another thing we need to do is to look for a cross-functional interaction of our metrics. That's just going across the company to say what is it that you need from me and maybe you're downstream from me in the company and I produce something that I hand off to you. Maybe it's information, maybe it's an assembly, and if I don't understand what it is that you need, then I’m going to make something and feel very happy about it and it may not in any way meet your requirements. Similarly if there are some things that I need to take a hit on, if I need to have a higher cost associated with let's say a raw material supplier or higher cost associated with maintenance in order to support the overall organization, I need to know that. Otherwise I’m going to make decisions in my area that hurt the company as a whole. So they need to be cross-functional as well.
The last one I think is really important: there needs to be agency in those who are responsible for the metric. By that I mean they need to be a part of the process. If I as an external consultant came to your company and started telling everybody what to measure, I might even be right about what to measure but nobody's going to listen to me. So I need to involve the people in the creation of their own metrics. Because if you think about it, if I’m part of the creation of my own metric I’m far more likely to think it's important, to think it's the best possible metric, and far more likely to actually follow and track those metrics in order to create the metrics that I’m going to then be using over time. So I think it's really important to involve the experts themselves in the creation of their own metrics.
Okay so I promised you one weird trick, because that's kind of the internet meme. It's not a weird trick but I thought that might be fun and it's going to focus on three different areas: one is the concept of emergence, another of piloting, and another the scientific method in business.
I’m going to show you this process to make metrics, but I’m going to hit a couple of different points before then.
Often I hear a concern that the problem is that there's too much to measure in my company or my organization, that there's millions of potential metrics and that's very true. But the way that the world actually works is there's a million things going on at any moment but not all of them are equally important. In fact, if you've ever heard of the Pareto principle before, you would know that the vast majority of variability - eighty percent of the variability - is controlled by 20 percent of, in this case, the metrics. And so we don't need to measure everything in the company for management purposes, we just need to measure the important things, and only a few of them are important.
The example here is if I was looking for a needle in a haystack, this is the needle that I’m looking for. I don't care about any small needles here; I want to find the big needles and I want to use those as levers in controlling the operation of my business. The tricky bit is figuring out which ones are the important metrics, no doubt about that.
Now these metrics do need to link up to what the organization is trying to do. There's lots of other things that I could measure that don't have any relationship to it, and I don't really care about those with the exception sometimes there's regulations that are required and so for government regulation we have to measure this thing. Well then it doesn't necessarily align with everything else in the business, but we still have to do it. But vast majority of what we're looking for here are metrics that are important that really control the activity of the overall accomplishments of the business itself.
Now another thing I want to spend a little time on, and I spend more time on this in the book, is the purpose of management because I think a lot of managers have a mistaken impression as to what their role actually is in the business. A big point right here management does not create customer value, the customer value is created by your front line employees, the ones who are providing the service, the ones who are making the raw material, the ones who are making the components, whatever. The people are who are doing the process that makes you money are the value creators. The value of management is how well they support those value creators in doing their day-to-day job, so it's an important function no doubt. But if a customer could get high quality product delivered on time without any managers, they wouldn't care. The managers are valuable only to the degree to which they can help those workers provide the value both in the short term, the tactical term, as well as the long term or the strategic term.
Now in order to be able to do that effectively, managers need metrics. I have a lot of sympathy for managers who are trying to do their job in the absence of having anything measurable about how things are going. It would be a terrifying position to be in. So what we're trying to do is to give managers at every level the numbers the metrics that they need in order to understand when they need to interfere with the process and support their people and when they can take a step back and let it run on its own, And of course, these metrics need to be related to the goals of the business, so the process I’m going to show you is going to try to answer all these different concerns and needs in one simple process.
Now I’m going to give you the one and only formula. You've got to have a little bit of math formula here, the one and only formula for the presentation, which is the level of your business success is a function of the value that's created at the front line and how well management supports that value creation at the front line.
So what we're trying to do is we're trying to build measures - metrics, throughout the entire organization so that management can do its job. So that management can support the people who are actually creating the things that we're able to sell.
Now there are other things that are enabled by the existence of management metrics. One is something you might call “leadership.” Higher level managers: part of their role is to lead the company. Leadership is defined as going in a direction or achieving something that they wouldn't have been able to do without something different, without that person who's setting that long-term objective, without addressing these very difficult problems. That's called strategic planning, and as you document your leadership it goes on to some sort of strategic plan that says these are the things that we need to change fundamentally from where we are now. We're not going to get there by working harder and smarter, we have to change the rules.
That's enabled by having the right metrics throughout the organization because then you go back up to the top of the organization and you look at the organization's top-level metrics to figure out which of those needs to make a step change - a big increase. You start pulling on that, and by understanding all the other metrics throughout the organization you can understand where to apply and to instantiate projects that will allow that big change, that big improvement to happen. We're not going to talk more about strategic planning, but that is enabled by the building of these metrics.
The other thing that's enabled is something that I would call daily management. That's the day-to-day activities that frontline managers do in order to support directly the front-line people who are actually doing the value creation. That looks different. Managing managers is different than managing the frontline or managing the process itself. There's a different process for that but I need that frontline manager to know what metrics are important before they can do their job very well and in order for them to figure out what their metrics are, I need the metrics for everybody else in the organization as well as the very top of the organization.
Okay so those are two topics we're not going to talk about. We probably don't have time, but those two things are enabled by having the right metrics in an organization.
So I talked about emergence. What is emergence? The idea here that we're using emergence for is how we manage a complicated entity like a business, like maybe a multinational business, with all these different things going on all the time. What we need is something that's going to encompass all this complexity to few simple rules, and that's exactly what I’m going to show you. If we follow simple rules everywhere in the business we'll be able to manage complexity, because we don't need a CEO who's got a massive brain to tell everybody what to do. We need everybody doing the right things without that level of intervention. This concept of emergence is that from simple rules comes complexity and is going to be exemplified by what I show you here.
We would start off saying for business, or for any other organization from non-profits as well, what do we want to be tell me 10 years from now? What is your organization going to look like, and if you can describe that in words? “We're the number one producer of this,” “we are known as good for that.” Whatever those things are that you want to become that you aren't currently starts telling us where we want to be. Once you have those words, you could call that a vision in our terminology, a vision is a 10 15 maybe 20-year word description of where the company's going to be in that time frame.
At the same time, we have to talk about the value that you're creating. Your organization exists to provide value to somebody. If it's a non-profit, you exist to turn donations and resources into services and things that people need. If you're for-profit, you exist to turn information and resources into things that people are willing to buy. Right, whatever that is that they're buying, whatever it is that they're wanting to get from you is the value proposition, is the value that you propose to give to your customers. We need to understand that and so we write those down as well, because if you don't understand the value that you're creating it's unlikely that your customers do. By the way, those of you that might be in marketing, if you understand your value proposition that this allows you to go into a market and start segregating the market based on who's looking for the thing that you're actually producing. And just as importantly, saying “no” when it's something that you don't do and you'd have to build an entirely different system to get them what it is that they want.
Well this is not infrequently done at companies: you have a bunch of words saying you know this is our vision this is our value proposition but if you stop there it's just a bunch of words. It doesn't tell me what to do. So I need to somehow turn those words into something that we can all objectively agree is what's going on, is what we're measuring, so we ask the question, “How do we measure that? How do we measure this thing we're calling a vision? What metrics does that generate this value that we say we're generating? How do we measure that? How do we know if we're achieving that? How does our customer know if we're achieving that?” When we answer these questions, we end up with top-level metrics, right? This is organizational success as defined by numbers, objectively defined success for the organization.
Notice that these top-level metrics directly come from what we say we want to be and what we say we create that is of value. Once we've defined these metrics, we can begin to gather data to see where we are versus where we'd like to be in order to achieve our longer-term vision.
Now of course this gives the top level of the organization success metrics, maybe a CEO, but it still doesn't tell me what to do. If I’m working on a shift and I’m working on a machine, I don't know how what I do fits into any of that stuff, right? I’m talking about profitability I’m talking about maybe some high-level success metrics in terms of market penetration. I don't get any of that stuff. I need to know what I need to do to support that. That doesn't tell me the answer to that, so the very next obvious question is to say, “All right, so how does everybody else contribute to that?”
This is where emergence comes in, because there's a very simple process that you're going to use in order to answer this question. The CEO, the president now has their top-level metrics of success and then the CEO turns to their direct reports, maybe a chief operating officer, maybe a chief information officer, maybe a chief financial officer, and says, “Here's what I need. These are the top-level metrics. This is what I need, what do you do that relates to that?” Notice how we are vastly decreasing the question. We're not asking a CFO “What can you measure?” we're asking a CFO, “What can you measure that relates to what the business is trying to accomplish: our value proposition and our vision?” And as they answer those questions, and every employee should be able to answer such questions: “Here's what I do that relates to that,” then we can measure that. “I contribute to profit by doing this.” Okay, let's measure that.
Notice how each level plays a role in the creation of their own metrics. At this level maybe we're talking about a chief operating officer who talks about scrap rates, he talks about production rates, he talks about all sorts of things in the production world. The reason that they're answering those particular metrics is because they're related to those top-level metrics.
Now as the COO does that they come up with their answer to, “If this is what my boss needs, what do I do that leads to that?” They do the exact same thing down to the next level. And so they may gather all their direct reports and say, “In order to get the CEO their stuff I’m measuring these seven or eight metrics. What do you do in your areas that relates to or supports me getting my metrics?” And we have a discussion and each individual can come up with a proposition as to what they think that they can measure in their area that relates to that and then we've created another level of metrics.
You can keep going down into the organization to get to that last level of management, and if you've done it correctly - if you've answered the question, “If my boss needs that, what do I need to measure to get him or her that?” By the time you get down to the front line, you've got linkage all the way from your first level of management all the way up to the top of your organization. You've got a linkage of all the things that we think that we do that relates to achieving the vision and the value proposition.
One observation here is err on the side of the person coming up with the metric. One of my favorite stories about this was somebody that I saw who had been at a very small company - it was actually an aluminum smelter - that was put into mothball. They'd turned off the smelting capacity and then a lawyer ended up buying it for a second mortgage on his house and restarted this because that was kind of the only reason that that that little town existed was to have workers for this this factory. A whole bunch of unemployment lawyers saw that as a need he bought this company and started it back up again.
So they brought somebody in early on, they started with maybe 10 different people. They brought somebody in as a secretary, but as the company started to grow and they started making really good business decisions they needed somebody to move up with them, and so the secretary was promoted and promoted and promoted. Finally, she was the one in charge of HR.
So she came to her human resources job in a very non-traditional way, but knew a lot of people in that company. When I sat down with her to create her metrics, one of the first things she wanted to measure was the “number of people who come into my office smiling each day.”
Now as an engineer, as a statistician, there are a lot of problems with that as a metric and I was a bit resistant at first, I must say. But she really was insistent that she wanted to have that as a metric that she would track over time. So I said, “Sure let's do it!” She had this little pull out part of her desk and a little piece of paper on there and every time somebody walked in, she'd put a tick mark if they're smiling or a dash mark if they weren't. And interestingly enough, as we tracked that data over time, the percent of people who were coming into her office smiling or not that turned out to be the best leading predictor of turnover six months down the road.
So I now tend to err on the side of the people who are creating the metrics, because, you know, they're the experts. She was the expert in HR in that company and having me judge her metrics creation because it was a little kind of crazy to me just meant that I didn't know her job very well and she did.
You might end up with some kind of strange metrics, and maybe you're willing to tolerate that for a bit because as I’m going to show you, we're going to come back and check with data to validate the metrics that we end up with. Because at this point, it's just kind of everybody's best guess as to what the metrics are for the area. If you haven't gathered any information yet, you may not even know what the best metrics are.
So you keep doing this at each level. Again, if you've done this correctly, then you should have a relationship at each level between what the people who are at that level and how that rolls up to their boss, how that rolls up to their boss, all the way until we get to the top of the organization. We've got alignment, at least as best as we can tell at this point without any data, to achieve the objectives of the company.
Once you have such a decision support system right up and down throughout the organization, managers at all levels can start making decisions based on data. And because the metrics are specifically those metrics that are related to achieving the business objectives: the long-term vision as well as value proposition, then you know that they're making the decisions as best as they can with the information that they have. You don't have to worry about them, you can give them the responsibility to make those decisions because the metrics they're basing those decisions on are related back up to the company.
A lot of fear in organizations right now is because people don't trust their managers to make the decisions they need to make, because they're making a different decision than I would make. Oftentimes that's because they either don't have any data, they don't have any metrics that are related to anything, or the metrics that they are basing their decisions on aren't related up into what the business needs them to do. So this is a way of giving that responsibility down to each level of management. As they create their own metrics, they're taking responsibility for their own metrics, for their own reactions.
And as I said, when you get down to the front line, you've got a front-line manager who's making decisions based on the metrics they have and how they're related up. But they're facing the process, the individual contributors, the employees who are creating the value of that organization. So as they do that, they can now work with the process and look for ways of maybe continuous autonomous local improvement to the process because they can specifically focus using their metrics on what's important for the company to get what it wants.
And this is how you can have local continuous improvement by the front line people but keep them on the rails as to what the company itself needs them to do. Another example I’ve seen of not that is at the same aluminum smelter. They had people who spent a lot of money on local improvement projects like making a calendar or like changing the chair that the people were sitting in. It wasn't really aligned with the business needs. You could take that same amount of talent and passion redeploy it and put it towards things that the organization needs to accomplish and have a lot of little improvements all over the place adding up to one really big improvement.
All right, so at this point we've got what I would call a decision support system built. Now at this point it may just be a bunch of words on a piece of paper, it may just be metrics that we think we want to do but we don't actually have access to do at this point. That's okay, because that's going to lead into the next step. Remember, I said this is a craft. We're going to have to work with this and we're going to have to learn it.
So that moves into the next phase - pilot testing. Kind of like this woman is doing here. She's going to test the water with her toe before jumping in all the way. Exactly the same thing is what we're going to do in an organization that hasn't had a lot of data to make decisions before, is we might even for the short period of time collect manual data, maybe write it down on a piece of paper or maybe type it into a spreadsheet. Collect it that way just for a little bit of time to find out if it is valuable. If you're collecting the data and it's just irritating and you never use it, get rid of it find something that is more useful. If as you're entering the data you're like, “Wow this is actually helping me make good management decisions!” maybe it's time to find a more effective way of tracking it over time.
This is how you drive the data science as opposed to the data scientists coming in saying “Tell us what to measure,” and then coming up with stuff, you actually go to the data scientist and say, “These are the things we need to figure out an easier way of measuring now that we know that it's useful, now that I’ve proven to myself that using data to make decisions is actually helpful.” Now you go to your computer scientists and your data scientists and have them build the data structures necessary for you to get the information faster and easier than ever before because you know it's worth it.
Additionally, even just the manual collection of data - writing it down on a piece of paper, entering it maybe into some sort of spreadsheet - builds the habit in your managers of using data to make decisions. Because as you gather that information, keep in mind these are metrics that they themselves have created so they kind of think it's worth the time to do. They themselves have created these metrics, they gather them over time, they're getting used to making decisions with data as opposed to making decisions in the spur of the moment or using their gut. And we want that discipline to be instilled in the people who are making decisions, so that as they get more and better and higher quality data, they can use that higher quality data to make better decisions.
So don't be afraid to pilot test this stuff, try it out see what works. And you'll find that a lot of the metrics turned out not to be okay - get rid of them. And you find that you're missing something - look up for it. That's the constantly working out and learning and getting better at this over time so it's never dead. It's never done and then we walk away. It's always learning something new, always finding something else that we that we want to work on.
And that's using the scientific method in business. The scientific method is a very powerful way of learning new information, so what we're doing is we're gathering information. It's provisional - we think it's important, but we don't know. We haven't tested it before. We then test our assumptions and see if it is in fact important and we learn from that.
Some people I’ve worked with really thought that this metric that they created was going to be important. They gather it, they do a correlation against a higher-level metric, and they find out that it has no effect. And they feel like that's a bad thing. It's not - it's a good thing, because what you've done is you've learned from that experience. You've learned that that thing that you thought was important turns out not to have been important at all. And that's a point which you can then seek new knowledge and find out what actually is important.
By looking at the analyses of these data, you can figure out if you're missing important metrics. You can look at the metrics that you have and if it's not explaining what you see in the next level, I’m missing something important. Maybe I’m measuring something that isn't important after all - I can get rid of it. Do I understand how strongly the metrics at my level control the metrics up at the next level? And if I understand that this one metric is a really strong indicator of what's going to happen to my boss, I probably want to really understand that metric and put controls around it so that I can get my boss what it is that they need.
This iterative process of examining the data, examining my premises, my assumptions, testing them with data, allows you to learn over time and end up with a better and better set of metrics.
Also, you know in the internet today they say the last something that you'll need. These are not the last metrics that you're going to need. These metrics evolve over time like the butterfly who starts off as a caterpillar goes into its chrysalis and emerges a beautiful butterfly. Expect your metrics to do the same thing.
In my experience, about 80 percent of the metrics that you come up with, if you're pretty good and I know a lot of you are, 80 percent of them will be right, 20 of them will be wrong. Either you missed something that's going to turn out to be important or you thought something was important, but it wasn't. But the point is that you've learned this, right? You're learning more about what aspects of your business are important in order to achieve those top-level requirements - the vision, the value proposition.
It's a simple process. It's only based on the question of “if my boss needs this, what do I need to do to get my boss that?” It's a simple process, but as my aikido sensei says, “It's simple but it's not easy.” It takes a lot of work, takes a lot of commitment to examining the data over time, to learning from what you're finding. to changing the data getting better metrics over time. getting the discipline of using data to make your decisions. But as you do that, you're going to be bypassing your competitors by far. Because my experience with businesses is that very few people do this.
So that's the process and I was told we have about 15 minutes’ worth of questions. I love getting questions, and I encourage you to ask questions either now or perhaps later and if you're interested in this type of thing. There are all sorts of ways we can stay connected. I write a sometimes-weekly blog and do podcasts as well, and for those of you that are speaking English as a second language, I speak really fast so maybe it's a good way to practice your English, or maybe it's just frustrating I don't know. but we are now open for any questions that you might have about all the stuff that we talked about.
FACILIITATOR: Hi! thank you so much that was a great presentation. We do have some questions in the comments, let me show you.
All right, the first one is from Melanie.
Question 1: How could we implement a complete program of metrics aligned with the company goals without increasing so much the cost to a small business? And then she said, “Or are these metrics just for big companies?”
Steve: That's a great question Melanie. In my experience, it's even more important to get good metrics for small companies. Big companies can be pretty wasteful. They can throw a lot of money at things and get through it because they've got a lot of money to spare. Small companies, especially startups, don't have the resources to throw at things that aren't important, that don't directly support their company objectives. So the way you would do this for a small company is exactly the same: What are you trying to accomplish? Where are you trying to go? How do you measure those things? What role does everybody play in doing this?
My perspective is this is even more important for small companies. Again, this the process is simple, right? The idea is simple, it doesn't have to take a lot of time or effort to do it in theory. It's only a couple of hours conversation to come up with metrics. In practice, it's much more difficult than that, but if you're willing to devote some time to it, the company size doesn't matter. In fact, think about it. The conversation that happens at each level is essentially the same no matter what size the business is. If you've got 15 different levels of management, there's still a couple hour conversation between each level of management. If you're only four levels or two levels or one level of management, there's a much shorter conversation time. So it's probably easier for you as well. I think it's even more important for small companies.
FACILITATOR: Let me see uh we have one more. Jhony said, ”Excellent tips to help in the daily life of the company. This improvement helps the board to achieve long-term goals.”
Steve: Yeah exactly! Because you start understanding where it is you're trying to go. The way I talk about this oftentimes is as if you're planning a vacation. So let's say you're going to go to Florida for a vacation. You know where you end up, and so whether you're going to take an airplane or you are going to drive there, as you’re doing that, you can make adjustments on the way, but you know you're going to Boca Raton Florida, right? That's where you're going to take your vacation.
But if you don't know where you're going, if sometimes you're going to Boca Raton and sometimes you're going to Sao Paulo, then it's difficult to figure out where you're going, what I’m supposed to be doing at this point. So that's exactly it: the high level can then say, “What we need to do is we need to make this improvement over this amount of time and we're going to stay focused on that.” And as we're staying focused on that, that's going to generate the projects that are necessary in order for you to achieve those improvements while not doing things that are good but they aren't associated with those improvements that you're trying to achieve. Great, Great observation.
FACILITATOR: Awesome! If anyone has any more questions and did not get the chance to send it here, please do follow him up on Facebook LinkedIn or the podcast and thank you so much Steve! It was a pleasure having you here. Thank you for contacting with us and everything!
Steve: Okay thanks everybody. I appreciate the opportunity and I hope I didn't talk too fast! I tend to do that because it is an exciting topic!
FACILIATATOR: We'll have this on live for a little while and some of the ones they have translations, so you'll be good even for that.
Steve: All right take care everyone! Thank you!
FACILLITATOR: Okay awesome have a good one!
