The Fuqua School of Business
100 Fuqua Drive
P.O. Box 90120
Durham, NC 27708-0125
Like this page
Researcher: Long Lines Hurt Sales
June 25, 2013
There are very few people who would tell you they like to stand in line. However, recent research from Duke University's Fuqua School of Business and Columbia Business School using video analysis from Scopix Solutions proves just how much retailers stand to lose when customers perceive a long wait.
Fuqua marketing professor Andres Musalem was part of a team that analyzed video from a supermarket in Latin America to determine how customers reacted when they spotted long lines. The results are in the paper "Measuring the Effect of Queues on Customer Purchases," published in the journal Management Science. The full paper can be found here.
Professor Musalem explains more about the findings in a Fuqua Q and A.
1. What made you want to study this particular area of retail?
Retailers need to make many different decisions, such as figuring out the right products to offer, pricing, whether to promote products, and the level of service to give to customers. Typically, good decisions come from good information. When thinking about prices, promotions, and assortments retailers have access to large amounts of data.
However, when thinking about customer service levels, the picture is remarkably different. For example, most of the time retailers don't know how long each customer has waited. We may have information about the customers' purchases, but not about the number of minutes that he or she might have been standing in line. Without this information, it becomes very difficult to understand whether or not customers are waiting too long and therefore make any recommendations about whether the service should be improved.
2. Tell us briefly how you used video to determine how customers would react to lines.
The very unique opportunity was getting access to data collected via digital cameras. Scopix Solutions placed the cameras inside a large supermarket. These cameras captured short videos every 30 minutes used to measure the length of lines. Combining this information with loyalty card data that tracks the purchases of customers over several months, we were able to study how waiting in line affects customers' buying behavior, which is a crucial piece of information to determine the right service level offered to customers.
A key decision was to select a section of the store to be tracked. One could intuitively consider the check out section as the most obvious unit of analysis. However, think of a customer with a shopping cart full of groceries that is ready to join a checkout line. That customer is probably already committed to go through the line and might not care about the amount of time needed to wait before paying. With that in mind, we considered the deli counter because customers clearly have an easier time choosing not to buy deli products. They can always choose to leave the store without buying anything during that visit, perhaps buy on another day, or get prepackaged products instead of fresh ones. Using videos we measured the number of people waiting in line and the number of employees working behind the counter, which is a measure of the speed of service. We were then able to relate these service levels to data on customer purchases, with the goal of understanding how changes in the waiting experience affect buying behavior.
3. Based on this data, what makes consumers join a line and buy a product?
Intuitively, a customer's decision to join a line depends on the number of people waiting and how fast it moves. In terms of the latter, this is related to how many employees are working behind the counter. For example, a line with 10 customers and one employee leads to much longer waiting times than a line with 10 customers and five employees. What we found is that customers react much more to the length of the line, than to the number of employees working behind the counter.
4. What implications does this have for retailers?
This actually has important implications that may run contrary to the common wisdom about how to design a waiting system. Two very different options are the pooled system and the split system. The pooled system is one where all customers join a single line which is served by multiple employees. The split system is one where customers join different lines, each of which is served by a different employee. This last system (split) has typically been regarded as a less efficient option, since there might be times in which one of the multiple lines is empty and therefore a free employee loses the opportunity to serve customers from other lines (if customers are not willing or able to switch lines).
This frequent recommendation of favoring the pooled system instead of the split one is extremely intuitive. However, in many contexts the pooled system may not be the best option. This is particularly relevant when customers have a choice and can therefore decide not join a line and hence not make a purchase. In fact, a pooled system which makes all consumers join the same line, leads to a single long line rather than multiple short ones. This single long line may in turn make customers less likely to join and make a purchase.
5. Should all retailers stay away from combining lines into a single one?
Not really. What we argue is that if a retailer switches from multiple lines to a single one, the retailer may need to provide information to customers explaining that even though the single line may look longer, it moves much faster.
6. Your results indicate that if the line gets longer, regardless of the number of employees serving it, sales may drop. How large is this impact?
The answer changes depending on length. We found that if the line is sufficiently short (0-9 customers) there is a very small impact on whether other customers might choose to join. On the other hand, a change from 10 to 15 customers in line, for example, may reduce the chances of another customer joining that line by roughly 30%.
7. Are all customers affected equally by the length of lines?
Clearly not. Some customers (typically the more price sensitive ones) have a much weaker reaction to waiting times. Other customers (the less price sensitive ones) have a much stronger reaction. In simple terms, we can think of two extreme segments: some customers care much more about prices while others care much more about the waiting experience.
8. What does this research mean for retailer promotions?
Suppose that the retailer decides to run a promotion lowering the price of an inexpensive product. This product is appealing to price sensitive customers. Therefore we will see more demand for this inexpensive product, making the line longer. Now consider the price insensitive customers who are interested in a more expensive product. They don't care much about prices, but they do care about waiting. Therefore, with a longer line the retailer risks losing the customers who prefer expensive products and shorter lines.
What we see is a case where we start with two products that are apparently independent from each other. However, the fact that they are sold behind the same counter combined with the existence of different customer segments, makes these two products highly interrelated. So, the moral of the story is that retailers need to be very careful when setting prices for products that may seem independent. By virtue of being sold in the same section these products have strong interrelations with each other.
Musalem conducted the research with Yina Lu and Marcelo Olivares from Columbia Business School and Ariel Schilkrut from Scopix Solutions.