Managers have long said they avoid changing prices for fear of “antagonizing customers.” Eric Anderson and Duncan Simester found no evidence that anyone had measured it, so they ran a randomized experiment and followed the customers for 28 months.

The experiment

A mid-sized publishing retailer, selling about 450 of its own products by catalog, sent 55,047 customers one of two versions of a catalog, at random. Thirty-six items were discounted in both. In one version the discounts averaged 34%; in the other, 62%. The group to watch had bought one of the 36 items in the three months before, at a price above the smaller discount.

Those customers, when they received the version with the deeper discount, placed 14.8% fewer orders over the next 28 months, about $90 less revenue each. The share who placed no order at all was 34.0%, against 27.1% with the smaller discounts. The authors called it a boycott Published.

PublishedAnderson and Simester, “Price Stickiness and Customer Antagonism,” Quarterly Journal of Economics, 2010. A second study at an apparel retailer with about 110,000 customers found a smaller effect: 2.4% fewer orders.

The effect was largest among the firm’s most valuable customers, “those whose prior purchases were most recent and at the highest prices” Published. After the study, both retailers stopped sending discount catalogs to customers who had recently bought the discounted items.

What it means for a price rise

The experiment is about a price going down, so read it for the mechanism. A customer who has bought from you carries a reference price, what they paid last time, and judges every later price against it. A new visitor has no reference. A price change is only a change to people who’ve bought before, and the most recent, highest-paying ones feel it most.

That shows up in three places when you raise prices:

• The rise itself. Repeat buyers compare it with their last receipt. A rise that tests well on new traffic can still cost you regulars.
• The discounts around it. A welcome discount beside the new price, or a sale a month later, shows a lower price to customers who just paid the higher one: the experiment exactly (chapter 9).
• The next markdown. If a raised price comes down, the people who paid it are the ones the catalog lost (chapter 10).

A stranger sees a price. A customer sees a change. Plan the rise for the customer.

Loyal doesn’t mean price-blind

The usual intuition is that loyal customers are the least price-sensitive. For small changes, habit may keep them buying. But the catalog’s best customers weren’t reacting to a price level. They were reacting to being treated worse than someone else, and many stopped buying across categories. My reading: loyal customers forgive a higher price more readily than an unfair one. The next chapter is about what “unfair” means.

Do this

• Pull two numbers: the share of last quarter’s revenue from repeat customers, and the number of customers who bought each SKU you plan to raise in the last 90 days. The second list is who hears first in chapter 8.