In this session, we shall talk more about how do you actually formally measure price elasticity. Earlier we discussed drivers of price sensitivity. What are the situations where consumers are more or less price sensitive? Now let's talk about how do you actually go about measuring price elasticity. So let's take note of one thing, earlier I was using the word price sensitivity, now I'm going to use the word elasticity. So why am I using the word elasticity, and how do we define elasticity? So elasticity is defined in a very specific way, is defined as what is your percentage change in demand when you change your price by one percent. So it looks very complicated. So the first question you can ask is why worry about percentage change in price and percentage change in demand? Why not just talk about change in demand divided by change in price? That could also be defined as elasticity, but we don't do that. Why do we use percentage change in demand and percentage change in price? So there are two advantages of doing this. There are others too but let's focus on those two. The first important benefit is, let's say you are looking at soap. If you measured the quantity of soap in bars of soap, you'll get one number, if you measured it in tons of soap you will get another number. So if you just use change in demand and change in price, you'll get a different answer depending on whether you're demand is measured in tons or in bars of soap. Same way if your price is measured in dollars versus euros you'd get a different answer. We don't want that. We want the measure of elasticity to be not dependent on the units in which we measure demand or the units in which we measure price. So if you use percentage, it's unit less, if you use percentage demand, is also unit less. So that's one advantage. Second advantage is because it is not specific to units, you can also make good comparisons across industries and across firms. You can say a demand for gasoline is more or less elastic than say demand for soap or demand for orange juice. So you can make comparisons across industries because it's not specific to a particular unit in terms of tons or dollar. So across industries comparisons become easier and more meaningful. Now that we've defined elasticity, let's look at another concept which is elastic or inelastic. People often use the phrase demand for my product is elastic. Others might say demand for my product or my service is inelastic. What do we mean by it being elastic or inelastic? What we're really talking about is the effect of price changes on revenues. Let's say I were to increase my price by one percent. If I increase my price by one percent, the question we ask ourselves is what will happen to revenues? If you're demand is very elastic, then you increase your price your unit sales will go down quite a bit, your revenues will actually go down. If your demand is inelastic and you increase your price by one percent, then your unit sales are now going to go down as much and your revenue will increase. So let's look at in specific terms. I increase my price by one percent, because I increase my price, let's say my unit sales go down by two percent. What will happen to my revenues? They will go down. Why? Because I increase my price by one percent, but unit sales went down by two percent. Consider another condition, where I increase my price by one percent, but unit sales only went down by half a percent. Now my revenues will go up. Why? Because the increase in price was one percent, the unit sales will only decreasing by half a percent. So my revenues will go up. So what is the break even point? The break even point is going to be one. Which means if my demand goes down by exactly one percent when my price goes up by one percent, then my revenues will not change. But if my demand goes down by more than one percent, then my revenues are going to go down. Now let's talk a little bit about how we measure price elasticity more formally. Now there are thousands of ways to measure elasticity. How do we put some structure on this? The way to put some structure on this is ask ourselves what are we measuring and then how are we measuring it? So we could be measuring actual purchases, focusing on what people actually are buying, or we could be focusing on people's intentions to buy. On the other side we looked at conditions of measurement. Are we measuring price elasticity in natural setting where consumers habitually buy? Or are we controlling the setting using some experimental methods? So if you combine the variables we measure along with conditions of measurement we broadly get four different ways to measure price elasticity. So let's look at the very first one on the top left. Actual purchases in a natural setting. This is what we call as sales data. We can get sales data from our products that are already on the market. We can get sales and price data and then we can do something with it to figure out elasticity. We'll talk more about it. Alternatively, we could in a very natural setting conduct a survey and ask people about their intentions to buy. We could conduct some experiments. Then finally we could do some even fancier analysis like trade-off analysis or conjoint analysis. By the way conjoint analysis was developed at the Wharton School by one of our senior colleagues Paul Green along with Professor Srinivasan. So it's a Wharton homegrown product and we'll talk about this a little bit. So let's look at each of these boxes in a little bit more detail. Let's look at the example where we conduct surveys. Let say I have a new online news surveys and I want to estimate price elasticity. So one of the things I could do is I could look at 600 representative potential users of this product. I could split them into six sub-samples of 100 each, and to each of these sub-samples I describe in some detail what this new product is and then ask them whether they'll be buying this or not or using it or not. But the way I do it is, to each of these six groups, I give them a different price. So everything else about the product is described in exactly the same way, but to one group I say the price is $10 will you buy it, to another group I say the price is $11 will you buy it, and the next rule I say 15. Then for each of these groups I count the number of yeses and then I plot them on a graph. If you look at this graph, you count the number of yeses it's higher at tenth lowest at 15. I can look at this and I say, I have a demand function which relates price to demand, and from this I can compute elasticity. So that's one way. Now clearly as you listen to me you must have said, there are many weaknesses of this method. Yes there are. But before we look at weaknesses, let's look at some of the strengths. It's quick, it's not as expensive, and of course there are many weaknesses. One of the biggest weaknesses, people often don't do what they say they will do. This in marketing we call as the intentions behavior link. There are ways to strengthen the intentions behavior link. We're not going to cover this in this module, but there are definitely ways to do that. The method also does not explicitly account for comparative prices. There's another important weakness especially in a service setting or a repeat purchase setting. You ask the question, will you buy this product? You didn't ask the question, how often will you buy it? How long will you buy it? Because your total revenues are not going to depend on whether people are going to just try this surveys, but it's also going to depend on how long they are going to by it, how long they are going to use it, and as Professor Fader said, what is the customer lifetime value. So that's one method, another alternative could be, we launch this service in the marketplace, we observe it for a few weeks and we change prices during that time. When we change prices we observe different sales. So now we have hard sales data. Once we have these data, we could plot these data or we could estimate some statistical models to estimate price sensitivity. For example we could run a regression on sales against price and see what is the formal relationship, and from that we could estimate price sensitivity. We could conduct some field experiments. For example, we could change prices in one setting, compare it with another price in another setting, and look at the difference. So a very nice example of a field experiment is from a research paper published by one of our own faculty members at Wharton, Professor Steve Hoch and some of his colleagues. He did this experiment when he was at Chicago. They co-opted a very well-known retail store in Chicago called Dominick's. If you happen to be from that area, you may be familiar with this store. This store had 90 odd stores and they split them into three groups, about evenly. For one of the groups, they lowered prices by nine percent. For the second group, they increased prices by nine percent. The third group, they kept the same. The third group is often referred to in experiments as the control group. So what were the results? Well, the results were very intriguing. For the stores where they lowered the price by nine percent, the unit sales went up by three percent. The stores where they increase the price by nine percent, unit sales went down by four percent. So let me ask you, what was the price elasticity observed based on what we covered? We lowered the price by nine percent, sales went up by three percent. What was the definition of price elasticity? Percentage change in demand, divided by percentage change in price. Percentage change in price is nine percent, percentage change in demand, three percent, elasticity, minus one-third, 3 divided by 9. Why is it minus? Because when you are raising price, sales are going down. When you're lowering prices, sales are going up. So elasticity observed in the stores where we lowered the price by nine percent was one-third. What does that mean? Was the demand elastic or inelastic? The answer is it was inelastic. When we increase the price by nine percent, how much did sales go down by? They went down by four percent, from 100 base to 96. Again, you see 4 divided by 9 is less than one, which means when we raise the price by nine percent, unit sales went down by four percent. Again, your revenues would have gone up. So in both these cases, we observe that the demand is relatively inelastic. So what would be the conclusion from the grocery store manager's perspective? Well, one conclusion could be because my demand is inelastic, I can raise prices, my revenues will go up. Now, that looks like a nice conclusion, but you have to be a little bit careful. Why do you have to be a little bit careful? If consumers are able to compare prices more easily, then maybe these effects will no longer remain valid. So if another competitor store were to emphasize that Dominick's has raised prices, they highlight it, they make it more accessible to the consumer, make consumers more sensitive, then maybe some consumers may shift stores. But what do we find usually is consumers are quite sticky when it comes to the stores they buy from. Maybe that's one of the reasons you're seeing low price elasticity in these settings. So that's one way to measure elasticity, is to conduct a field experiment. The last method we're going to talk about is trade-off analysis or conjoint analysis. So I'll just give you a sense of what this method is to give you an idea of how this is done. Then of course, there are more formal ways of doing it. There are companies that specialize in conducting this. You may talk to one of those, or read more about it in a book. So let's look at the following example. Let's say we have six restaurants; A, B, C, D, and E, and F. We have, these restaurants are defined on two different characteristics or two different attributes. One is the quality of their food and other is the atmosphere or the ambiance. So the food quality can be either excellent, good, or fair and the ambiance is either intimate, candlelight, or bright lights. Now with these three food quality and two types of ambiance, we can create six different types of restaurants. So I've done this in a very purposeful way. Now let's say I ask one of the consumers or potential buyers, "Tell me which is your most preferred restaurant." Let's say this person says A is the most preferred restaurant. Then I ask them the next question, "Suppose A were not available, which is the next restaurant you will pick?" Let's say this person says B. Then I tell them, "If B is not available, which one will you pick?" They says C. In effect, what I'm trying to do is ask them to rank order these restaurants. Now let's look at another customer. Let's say, I ask them "Which is your most preferred restaurant?" They say A, I asked, "A is not available, which one do you pick? "They say C. I say, "That's not available, which one do you pick?" They say E. Now, if you look at A, B, C versus A, C, E, would you not agree with me that you learned something about these consumers? What did we learn? Consumer number one cares mostly about the food quality. Consumer number two cares much more about ambiance than food quality. Now this is just to give you an idea that we are able to judge from these choices, in a very purposeful way, does the customer care more about food quality or ambiance? We can then quantify it also more formally as to how much more do they care about food quality versus ambiance? Now, let's think about one of the variables instead of being food quality could be price. So then we can say, "How much does the person make a trade-off between price and food quality?" That will allow us to measure price sensitivity. So looking back, there are four broad ways of measuring price elasticity. These four broad ways come about from, what do we measure and how do we measure it. Are we measuring people's actual purchase behavior or are we measuring their intentions to buy? In terms of how we are measuring, are we measuring it in their natural setting, and then are we measuring it in a controlled experimental setting? So this two-by-two table that we discussed gives you four broad different ways. What we did in this module was to give examples of each one of those four different ways of measuring price elasticity.