EACH November for the past three years PayPal, an online-payments service owned by eBay, has published data showing the volume of payments it handles on what retailers call “Black Friday”, or the first day after the Thanksgiving holiday. The idea is to assess the strength of consumer demand on the day that marks the start of the holiday shopping season. On Black Friday last year the total of payments made on PayPal’s system was 20% higher than on the same day in 2008, suggesting that consumer confidence in America was looking more robust. Retailers are hoping for a repeat performance this month.

PayPal is one of a growing band of online companies dipping into the data they gather in an effort to divine trends in the American economy. Last month Hal Varian, Google’s chief economist, revealed that the search-engine giant had developed a “Google Price Index”, based on web-shopping data it holds—though it has yet to decide whether to publish it. While Google flirts with inflation, Intuit monitors employment. The firm, which offers online payroll, tax and other services to small businesses, produces a monthly small-business employment index based on aggregated data from 59,000 of its customers.

Web firms see such indicators as something of a sideline. Tayloe Stansbury, Intuit’s chief technology officer, says that most of the firm’s data mining is geared to helping its customers. But some economic policymakers are paying attention to web firms’ statistics, for a couple of reasons.

The first is the speed with which the data are generated and crunched. Ceridian, which manages payments made by businesses via the web or with stored-value cards, tracks aggregate purchases of diesel fuel by truck drivers in real time. Together with the UCLA Anderson School of Management, it has used these data to create an index to gauge what is happening to shipments of goods by road in America. The latest instalment, published this week, showed the index had fallen by 0.6% in October compared with the previous month. Given that October is usually a busy time for the trucking industry the index’s creators say the decline “sounds an alarm” for growth in the fourth quarter.

Because web businesses gather data rapidly, their indicators can sometimes identify trends before official statistics. Take the case of Monster Worldwide, an online job service that publishes an index tracking jobs posted on its own and other sites. This fell sharply in 2007 before official numbers showed employment in America weakening (see chart). Google has said its search data may also provide useful early-warning signals: it is exploring whether searches for terms such as “unemployment insurance” are a good way to predict future increases in joblessness.

A second reason that web firms’ indicators are gaining popularity is the detailed data that underpin them. John Krainer, an economist at the Federal Reserve Bank of San Francisco, says he has drawn on statistics from Zillow, an online property site, in his research because of their “granularity”. The Zillow Home Value Index draws on the site’s individual valuations of 72m houses across America.

Some economists caution that web firms’ data have big handicaps. Many of the indices have only a short history, which means they are of little value to policymakers interested in long-term trends. And they often measure only online transactions, which limits their appeal. Both caveats carry some weight. But as more economic activity moves online, the notion of using bits and bytes to measure booms and busts will surely become more attractive.