Articles

Choice Trumps Price on the Internet

Selection ranks above price among the benefits of shopping online

The New York Times, "Economic Scene" , April 21, 2004

What is so great about Internet commerce?

To many economists, the answer is simple: lower prices. With a world of suppliers available at the click of a mouse, buyers can easily find out who has the best deals. Online commerce looks something like the blackboard model of perfect competition.

"When I first started doing work on how the Internet is affecting commerce, like a lot of people, I was really excited by this nearly perfect market," said Erik Brynjolfsson of the Sloan School of Management at Massachusetts Institute of Technology.

His early research found that prices on the Internet were 6 percent to 16 percent lower than prices off-line.

But when he thought about how people actually shop online, and what they find valuable, he realized that low prices are not the big story. Selection is. The Internet offers variety that is simply impossible in traditional stores.

When I wanted a contemporary light fixture in copper, I used Google to find a specialty retailer that had one I liked. I recently did the same thing to find a particular brand of Velcro-sealed envelope that I use for receipts when I travel. I regularly turn to Amazon.com and Alibris for books I cannot find in local bookstores or even libraries.

Online shoppers are not just buying the same stuff for less money. They are buying different stuff. And they are much more likely to be getting exactly what they want than are off-line shoppers. Wal-Mart has low prices, but Walmart.com carries six times as many items as the largest Wal-Mart store, the article says. "Amazon's slogan is world's biggest selection, not world's cheapest prices," said Professor Brynjolfsson, who has done pioneering research on information technology and productivity.

All this variety could be overwhelming. But consumers do not have to sort through item by item. Online shopping includes tools like search engines and customer review sites, or Amazon's many referral services.

You are not only more likely to find what you are looking for online. You are more likely to discover something you like that you did not already know about, Professor Brynjolfsson said. Partly through links and referrals, the Internet increases sales of obscure products. In 1997 and 1998, in the early days of Internet commerce, The MIT Press reported 12 percent annual increases in sales of backlist books, thanks to Internet retailers.

"In effect, the emergence of online retailers places a specialty store and a personalized shopping assistant at every shopper's desk," write Professor Brynjolfsson, Yu Hu, and Michael D. Smith in a November 2003 article in Management Science. "This improves the welfare of these consumers by allowing them to locate and buy specialty products they otherwise would not have purchased due to high transaction costs or low product awareness." (The article, "Consumer Surplus in the Digital Economy: Estimating the Value of Increased Product Variety at Online Booksellers," is available here.)

In the article, the authors, all economists, estimate just how much better off consumers are because of the variety available online. They look specifically at "obscure titles," books that rank below the top 100,000 in Amazon.com sales and probably would not be carried in a traditional bookstore. (The typical Barnes & Noble or Borders superstore carries about 100,000 titles, while large independent bookstores stock about 40,000.)

Using Amazon rankings and publisher data on 324 titles, the researchers determined that nearly half the book sales at Amazon, 46 percent in 2000, were of obscure titles.

They then "tried to calculate what people would have been willing to pay for these books versus what they actually did pay," Professor Brynjolfsson explained. That's the concept economists call "consumer surplus." If you buy an ice cream cone for $2 but would have been willing to pay $5, you get $3 of consumer surplus.

By estimating what the demand curve for books looks like, using well-established techniques, the researchers could estimate the consumer surplus for all buyers in this market.

The results are striking. People are really happy to find obscure books, and would be willing to pay far more for them.

"The consumer surplus was about 70 percent of the purchase price for each book sold," Professor Brynjolfsson said. "If a book was purchased for $20 on average, consumers would have been willing to pay on average up to $34."

All those benefits add up to big money -- around $1 billion in 2000. By comparison, Amazon's lower prices saved consumers about $100 million that year.

"So they got about 10 times as much value from the selection as they got from the lower prices and competition," Professor Brynjolfsson said. "An order of magnitude more value was created from the increased choice and selection."

The same thing is almost certainly going on for goods like music CD's and DVD's. More speculatively, he suggests, we can imagine something similar happening with job services like Monster or even online dating markets.

The combination of more choices and easier searches creates big benefits for consumers.

"I won't endeavor to try to measure the value of those relationships, but certainly there are a lot of people finding matches who previously maybe wouldn't have found matches or wouldn't have found as good matches," he said.

Clearly, the implications of this research go far beyond Amazon or books. More and more economic value seems to be coming from giving consumers greater choice, off-line as well as online. Yet these intangible benefits, which represent real increases in the standard of living, are not picked up in most economic measures.

"If you spend the exact same dollars but you spend it over a broader set of choices, economic theory, and, to some extent, common sense, tells you you're getting more utility out of that -- if you can buy different flavors of ice cream instead of just one, or different colors and sizes of shirts," Professor Brynjolfsson said. "Our methodology tries to quantify that for one piece of the economy."