Okay, the second step in the process is actually setting targets for these non-financial measures. Now you hear a lot of times that more is better, 100% of your employees or customers need to be satisfied, which may be true. But in a lot of cases it may not be true. Again, it's not the same as a financial metric where more money and more profits is always preferred, right? The problem is the companies we talked to they have considerable difficulty setting goals for any of these non-financial performance measures and there's a number of reasons for this. One, what do you do when the measures have no common denominator, okay? This is not like financial results, I know how to do a budget because everything rolls up from the bottom to the top because everything's in common currency. How do I compare a 10% decrease in customer complaints to a 3% reduction in defect rates? They're completely different, so if I wanted to say which one makes more sense well I gotta put them in a common denominator which just happens to be currency. Which is why we're going to link this to financial performance because it's going to allow us to kind of make tradeoffs. The other big problem is, you got these non linear functional relationships and tradeoffs. What that means is, you can have either economies of scale or diseconomies of scale, right. More is not necessarily better. I probably do not want 100% satisfied customers. The reason is it's really expensive. And ultimately if what I'm trying to do is optimize financial performance, that may not be the right thing to do. Same thing with employee turnover. Well yeah employee turnover is very expensive. But it'd be expensive to keep people, and in truth a lot of companies actually want some turnover because it creates new opportunities for people to move up in the organization. It brings in new world views, so you probably don't want zero employee turnover. You probably don't want 100% satisfied customers. The question is where's this point of diminishing returns and how can we figure out where that's at? So let's go through an example of this. Okay, here's a personal computer manufacturer. It turns out the margins on a personal computer are very small. You do not make much money on each personal computer. The other thing with a personal computer is you don't necessarily buy one every year. So there could be a lag between me being satisfied and me deciding whether I'm going to buy from them again. On the other hand, if you don't like your personal computer, I'd bet you'd tell somebody. So you've got this whole issue with word of mouth. So here are some hypotheses in the business model of this PC manufacturer. One is more satisfied customers recommend the company's products to others. It's also called positive word of mouth. Dissatisfied customers complain to others about the product or negative word of mouth. Turns out they've done studies in the auto industry, if you really like your car, you tell one person. If you really do not like your car you tell 10. So it turns out negative word of mouth can be worse than the positive word of mouth and especially in a business like this where you do not necessarily buy a PC every year. The word-of-mouth may be that the direct relationship in a lot of cases between improving satisfaction and short-term financial performance. Okay? So here are the hypotheses we can test. Greater positive word-of-mouth increases future financial performance. Greater negative word-of-mouth reduces future financial performance. That's what they're hypothesizing. So if you believe that, what are you gonna do? Spend money to increase satisfaction because that should change your word of mouth. So here's what they did when they actually went back and looked at this. So what you have here, if you think about when you buy a personal computer, you have a warranty card. You fill that out. They kinda know who you are as long as you filled out the warranty card. And what you're trying to do with a lot of these non-financials, see if they are leading indicators of future financial performance. If I know something about satisfaction now, does it tell me something about future financial performance? Now remember, one of their hypotheses is if I have higher customer satisfaction, I'm gonna tell more people that I liked it, more positive word of mouth and vice versa. Well their satisfaction scale, like a lot of companies, it's on a five point scale. Five is I'm really satisfied, one is I'm not really satisfied. So when you do the analytics, here's what you find. First of all, we're gonna start off with a positive word of mouth. Obviously, if I don't like it, I don't give any positive word of mouth. If I don't like it, but not quite extremely bad, if I give it a two, which is that second red bar there, I might say something. After I get three, I'm kind of satisfied, from three to five. Yeah, I recommend it, but if you look at that, those lines aren't that much different. In fact, they go down a little bit if anything, but they're statistically the same. Basically, what this says is if I'm moderately satisfied right? I recommend it. After that, I could spend money making you more satisfied but you don't recommend any more than you would've otherwise. Now, here's why this is important. The company had a bonus plan. The bonus plan said maximize the number of people that give us a five for customer satisfaction. It's called a top box measure. What percentage of the people answer in the top box of the customer satisfaction scale? Very common bonus plan. Well if I was the manager and that's the bonus you gave me, all those people in three and four I would try to get them into five cuz you're going to pay me more. But looking at this, I'm spending money to get them into the five bucket but they're not recommending any more. It may not make it any economic sense. Let's go to the down side. Exactly opposite. Okay, if you look at negative word of mouth, obviously, if you don't like it you tell people. From three on you get the exact same answer here. So one of the big issues that comes out of this is, how do you want to set your performance targets? Now remember the target here was get everybody on five. Over to the right on the bar. That means get anybody on those threes and fours, push them up. Well what this may say is that's really not a very good target. Now I've asked a lot of students when I presented this, where would you set the target? And a lot of times they respond, oh why don't we set it at three, because that kind of seems like the hot spot. Well in truth, that's not right. What you really want to do is set the target such that nobody gets into one. That's where the biggest bang for the buck is. Get everybody out of the one category. Targets do not have to be set at the top of the scales. We seem to think that that's where it has to be. Get everybody up at the top. No, get everybody out of the bottom. And the only way you would have known that for setting this target is to actually do the analytics on it. Let's actually go back, test that prediction that if I have more satisfied customers. In the future they will recommend my product and that's gonna lead to better financial performance. Okay, now this need not be true so here's an opposite example. Okay, this is a company, a financial company where you're gonna invest your money with an investment advisor. Say you got your retirement fund and I wanna invest it there. What they wanted to look at was is there a relationship between satisfaction with our firm and future financial performance. Well, we did the analysis and guess what? We found absolutely zero. There was no relationship between satisfaction with the firm and firm performance. What you did find was that satisfaction with your investment advisor. This is where you have to figure out what the right measures are. It's not just I'm satisfied with the firm, right. I don't care about your research reports, I don't care about your offices. If I'm satisfied with my individual investment officer, then I tend to give you more money. Now the thing you care about in an industry like this is assets invested or what's also called assets under management. I want all your portfolio. Give it all to me because I make money on how much money I'm investing. So when you start looking at this, you see a relationship between investor, advisor satisfaction, and assets invested. Where they have an index that has many things in it. So here's low. The scale range is from one is I'm really not satisfied with my advisor. To seven I'm very satisfied. So here's the low end. Well It does start actually going up, right? As you're more satisfied you start investing more there. We're not seeing that diminishing returns that we saw with the computer manufacturer now though. Here's an example where you do wanna spend money to get them as satisfied as possible. There is a huge jump here in how satisfied you are with your investment advisor and assets under management. This is after controlling for how much money you made. So this just not they made a lot of money for you. This is one where you get exactly the opposite result from what we had with the personal computer manufacturer. There it was, no don't put people in the four's and five's because after three it makes no difference. Here's one where it's exactly opposite. You get a huge jump between six and seven. So investing, and trying to improve investor advisor satisfaction, would seem to make a huge difference here, as opposed to the other one, and that's where you need to do analytics. Don't let anybody tell you they know. What's gonna happen to these relationships without trying to test it? So here's an example of trying to do the same analysis using external market data. It turns out that there's something called the American Customer Satisfaction Index. What happens is the University of Michigan goes out and surveys customers on their use of various products. And this is all published in one of the major business publications. So this is gonna come out. The question is when this comes out, does the market actually respond to these satisfaction index? Right? Do they believe that if a company has very good satisfaction, they're gonna have future cash flows that are higher? If they're bad is it gonna be lower? Now, this is done in many countries, including many countries in Europe. So there is a lot of work on this. Now what we're showing here is, what happens to the valuation of your company depending on what your customer satisfaction level is? This is after controlling for your accounting book values. So is your company worth more than your accounting book values would say? Based on what kind of satisfaction you have. Well, again if you look at the bottom, of the quartile one. Yeah, if you have fairly low customer satisfaction based on this index, your market value is gonna be lower. It goes up when you go into quartile two, better satisfaction, but again, look at three and four. Even the stock market believes that more is not necessarily better. After you've kind of met, got into the top end of the customer satisfaction index the market is not rewarding you with higher market value based on your customer satisfaction scores on this. Now again if you look at what happened to your stock returns. When this was announced, did the markets say, okay, now I know something about your satisfaction, I'm going to hurt your stock price today or increase it. This is looking over 10 days and it's called excess returns because we're taking out what happened in the stock market and your industry and things like that. They really hammered you. If you were in the bottom of the customer satisfaction index, your stock price went down 2% over 10 days, which is pretty big. You look at two and three, it went up, but fairly small amounts. But here in terms of the stock returns, not the valuation, the market did reward you for being on the top end here. Now it turns out this actually differs depending on the industry. The one industry where you do not see this was retail. And in retail, what you've found was that in quartile four. The top one. It actually was negative. You say well oh that makes no sense. Until you start thinking about retail. Think of the high end retailers with very high customer satisfaction scores. In the United States it's something like Nordstroms. You can go in there and there's a guy playing the piano. I don't go in to stores because they're playing pianos but some people do. There's apocryphal stories of Nordstroms will take anything back. So they've taken back tires, they do not sell tires. They've got very satisfied customers. What the market believes though is they've gone a little crazy. They're just too satisfied, you're spending too much money and it's just not paying off. So again, what you wanna do is, even if you're gonna look externally, make sure you kinda slice and dice the analytics to see does this hold across everybody. Does it vary by location? Does it vary by country, does it vary by which industry you're in? Cuz it turns out if you've got Chinese speaking customers they answer on a different part of the customer satisfaction scale. So you need to control for that. So if I'm trying to pool Chinese customers and non Chinese customers your statistics are gonna show nothing. What you need to realize is, you gotta estimate separate models with Chinese-speaking people, and non-Chinese. It's not a bias. It's just you tend to answer on a different part of a scale.
