Faculty Q&As

Faculty report from the frontier of management

Duncan Simester

1. Your research focuses on understanding customer behavior and how firms respond to that behavior. Can you explain your most current studies?
We have a study right now on price stickiness, why a price is “sticky.” Macroeconomists care a lot about this because it is really the heart of monetary policy. If prices aren't sticky, then using interest rates to change the supply of money will simply lead to a change in prices. On the other hand, if prices are sticky, meaning that they are slow to adjust, then interest rate changes can lead to other more desirable outcomes. There are all sorts of theories about why prices are sticky, and one is that managers are reluctant to vary prices because they worry about antagonizing customers. We have a study to explore this.

2. What is the approach?
We took a very large sample of customers who had previously purchased from a firm and used a catalog to experimentally vary the prices that they saw on an item that they had previously purchased. We found that merely sending a catalog that contained lower prices than what customers had previously paid prompted customers to stop purchasing from the firm. The effect was limited to people who had previously purchased one of the items and was stronger if their prior purchases were more recent or at higher prices. The design of the study allows us to rule out a range of alternative explanations. It appears that varying prices did antagonize the firm's customers.

3. Why do people stop purchasing?
Studies have suggested that customers who paid higher prices in the past may experience regret for not waiting, or jealousy toward customers who got a better deal. A price reduction may also suggest that the firm could have profitably sold the item at a lower price, raising concerns about fairness. The effects appear to confirm that managers' concerns are well founded and are large enough to contribute to price stickiness. You can imagine that this impacts the firms, but it also impacts monetary policy. It is important.

4. Are you also looking at the impact of sales taxes?
Yes. As you may know, if you buy something over the Internet, you are not charged sales tax by the firm. This changes if the firm has a physical presence in the state. If the Internet or catalog firm has a store or even a distribution center in the state, because of this physical presence, the state can oblige the firm to pay sales taxes. This means there is really a discrepancy … you can imagine the firms that have a physical presence are very unhappy about it because sales taxes can add up to nine percent to the price of an item. We look at whether this affects customer behavior, and whether this affects firm behavior. First, does collecting sales tax have a negative impact on sales through Internet and catalog channels? If there is a negative impact on sales, then we might expect this to also affect firm behavior. This motivates our second research question: Are direct retailers less likely to establish a physical presence in high-tax states?

5. How do you study this?
We obtained transaction data from a firm that opened its first store in a state. We look at customers on either side of the state border. All of the customers live a long way from the store, too far to go to the store to shop. Customers on one side of the border are obliged to pay sales tax once the store opens, while customers on the other side of the border are not. By comparing their transactions before and after the store opening we can evaluate how the obligation to charge sales taxes affects purchases. You would expect that when a firm starts charging sales taxes, customers who are obliged to pay them will stop purchasing or at least reduce their purchases from the firm. That is partly true. If they are purchasing over the Internet, purchases go down about 16 percent. On the other hand, if they are purchasing through a catalog (by calling on the phone or mailing in an order form), there is no change in their purchases. This is a very robust result: there is a big drop in Internet demand, and no change in catalog demand.

6. Why?
We are fairly confident that it's the ease of search. If you an Internet customer, you can find competing products by going to Google and searching through different companies. If you are a catalog customer, you need to have the physical catalog. You just don't have the same availability of outside options. Why do we think this is true? Well, consider purchases of items that have been discounted. If an item is discounted there is less incentive to search for lower prices elsewhere. Even on the Internet, the obligation to pay sales taxes does not lead to a drop in demand for discounted items. The discount takes away the reason to search.

The second half of the paper looks at: How does this impact firms? In other words, do firms avoid opening stores in high-tax states? It turns out that firms that have very little catalog or Internet business actually favor opening stores in high tax states. It is not that they like the high taxes — instead the high taxes are presumably correlated with other factors that make the states attractive, such as a high concentration of wealthy customers. On the other hand, companies that conduct a large proportion of their business over the Internet or through catalogs systematically avoid opening their first store in high tax states.

7. What have you discovered about customer behavior that is the most surprising or contradictory to popular beliefs about marketing?
I think that customers are typically much more sensitive to price cues than they are to actual prices. Customers often lack good information about what the price of an item is, but they want to get a good deal and so they evaluate all sorts of cues to determine whether they get a good deal or not. Cues such as a sale sign or a “9” digit price ending. You can take the same shirt and charge $34, $39, or $44 for the shirt, and people will buy more when it is priced $39, versus $34 or $44. They use the “9” digit ending as a cue about the price.

8. Because it is perceived to be a discounted price?
Yes. When retailers discount apparel, they tend to give it a “9” digit ending. Another good example of how price cues work is the employee discount promotions that the automobile manufacturers had in 2005. That was really a very effective price cue. There was a huge increase in demand, even though the prices of many models actually increased because prior to the employee discounts, there were already very big incentives on the cars. Customers did not realize this. They were more sensitive to the employee discount cue than to the actual change in prices.

9. You have said that marketing policies that maximize short-term outcomes generally lead to poor long-term outcomes. Can you explain this?
We can think of two marketing actions that we are confident have a big long-term impact on customer behavior: prices and advertising. Firms generally measure the short-term outcomes of these. We have completed several where we have varied prices and shown that the long-run outcome is often the reverse of the short-run outcome.

With pricing, if firms offer deep discounts, customers purchase. But we have observed that the short-run increase in sales does not last for the long run. It is fairly well-known that your best customers often just forward buy. They stock up. There is a negative long-run outcome. On the other hand, new customers — who haven't purchased from you in the past — tend to respond to deep discounts by coming back more often in the future. In this case the long-run outcome is positive. This is because these new customers are still forming perceptions of how much value your firm offers. These perceptions are very sensitive to the outcome of those first transactions. In contrast, existing customers have had lots of transactions with you; they have seen a lot of your products and your prices. The deep discounts don't change their price expectations very much. With existing customers you merely prompt forward buying, whereas with the new customers, we see a very positive long-run effect. There are two very different outcomes, depending on the customer.

We can also look at advertising, which we again study using catalogs. When you mail catalogs to people, if they are your best customers, it tends to steal from future sales. You will see a big short-run lift in sales but there is a drop in subsequent orders. If firms only measure the short-run outcome, they will tend to over-mail to those customers. That's how some companies end up sending 50 catalogs a year to their best customers. If firms only measure the short-run outcome, they sabotage future sales. On the other hand, for customers who are not as good, there is a positive long-run outcome. If firms considered to the long-run effects they would increase their mailings to weaker customers but reduce mailings to their best customers.

10. Are companies doing that?
Some are. I've been a professor for 15 years. I see more firms understanding this issue now than I did 15 years ago. The very best companies are now trying to come up with means of capturing long-run effects. They recognize the need to look beyond the short-term, particularly when it comes to pricing and advertising.

11. What are more general ways in which the Internet has impacted customer expectations and behavior?
The answer varies across markets. In some markets, such as travel and books, a large proportion of customers use the Internet to learn about prices and product features. In other markets, such as grocery products and gasoline, the Internet has not affected customers' behavior.

In markets where customers do use the Internet to search for information, the outcome really depends on what customers are looking for, and what they learn. If the Internet tells customers more about differences between products, then customers tend to become less price-sensitive. Instead, they are willing to pay a premium for their favorite products. If the Internet tells customers that there are no differences between products, then they will become more price sensitive, particularly if they also get more information about prices.

12. How is the economy impacting customer perceptions and how should retailers change their marketing decisions to adjust for it?
We have two studies going now with two very big retailers looking at precisely this question. We are using the 2001 recession to look at how it affected prices and which product categories people purchased. This issue is obviously important to monetary policy, but I don't have any answers for you yet.

13. Where is marketing research headed?
I think we are seeing a more interdisciplinary approach to research. Marketing is a field that has a quantitative side to it—economics, operations research and statistics—and a behavioral side to it, the discipline of psychology. We are seeing a trend of more experimentation in the field by the economists and statisticians. This is a ripe opportunity for interdisciplinary work with psychologists. For example, we have a study now which looks at the impact of disgusting products. Some products evoke feelings of disgust, even on a subconscious level … cat litter, feminine hygiene products, and such. In a study of a very large national chain of small grocery stores, we use a list of the 50 most popular non-food categories in the store, and ask people to rate how disgusting they are. We then look at a sample of more than 10 million transactions from a chain of convenience stores and show that if customers place disgusting items in their shopping baskets, then they tend to buy fewer food items. This is true even though the disgusting items are obviously new and in their original packaging. If customers buy disgusting items they avoid buying food. It really changes their purchasing behavior. This obviously has important implications for the layout of the store and which products should be promoted together. It is also a good example of psychologists and economists working together.

Related links