Jupiter shoots for the moon

"SPS," as Jupiter employees and clients call it, is now Jupiter's bread and butter. SPS customers are charged an average of $29,500 (although the cost can rise to the hundreds of thousands) for a series of special 25-page reports focusing on different topics within categories like "Site Operation Strategies," "Online Advertising Strategies" or "Consumer Content Strategies." With the reports come special one-on-one consulting sessions with Jupiter analysts, discounted tickets to the 10 conferences (such as "Plug.In: The Jupiter Music Forum" and "Digital Kids: Marketing to the Post-Modern Kid") that Jupiter holds every year and invitations to special "executive breakfasts" across the globe, plus assorted other perks.

In early1997, Jupiter had 17 SPS clients; today, it boasts over 500 -- a drop in the bucket compared to Garner Group's 35,000, but a number that includes many heavy-hitting Internet companies like CNET, Time Warner and America Online. According to Jupiter's S-1 filings, 57 percent of the company's income now comes from SPS services, another 28 percent from its popular conferences and the rest from subscription newsletters, longer studies and customized research.

But just how accurate and useful those consulting sessions and research reports are is up for debate. Silicon Alley Reporter editor Jason McCabe Calacanis says, "They definitely know what they are talking about -- any time you talk to the most intelligent people in the industry as your job, you will become intelligent. Their observations are well worth the price that people pay for them."

Jupiter CEO DeRose guesses that his company's forecasts and figures are right on the money about 60 to 75 percent of the time. He happily hands out a promotional leaflet to prospective clients, titled "Making the Tough Calls," that touts predictions the company got right. Among these are the no-brainer 1992 forecast that "access to the Internet will become essential for online providers" and a 1994 prediction that "the Web will become the ideal platform for commerce" -- along with more credible statistics such as a 1994 prediction that 5.9 million teens would be online in 1998 (actual number: 7.2 million) and a 1995 prediction that online advertising revenue would be $2.2 billion by 1998 (actual number: $1.9 billion).

Jupiter says it uses a secret set of metrics and projection models for its forecasting, but analyst predictions aren't exactly a hard science. Jupiter founder Josh Harris describes the process of forecasting numbers as "really puzzling the market, plus voodoo -- a little black magic doesn't hurt." How do you know that the numbers you are coming up with are right? "You just know," he emphasizes. "The number pops in your head and you know. It's coded in your brain." And the company's S-1 admits that "Internet commerce is relatively new and is undergoing frequent and dramatic changes ... it is very difficult to provide predictions and projections as to the future marketplace, revenue models, and competitive factors."

A former Jupiter analyst points out another flaw in the system: Since Jupiter bases much of its research on conversations with Internet companies, its data is subject to each company's own hype. "When you meet with people from companies, who demo their products and talk about the future, the companies are often accompanied by a publicist," says the analyst, who requested anonymity. "What they are telling you may be a lot of spin. To a certain degree you have to be willing to go out and find out how much of it is whitewash, and how much is true ... Good reporting and a little bit of skepticism would have done people a lot of good."

This, perhaps, explains why different market research firms can come up with different numbers and predictions in their reports: While Jupiter last year offered the statistic that online commerce would hit $41 billion in 2002, for example, Forrester gave the more robust numbers of $108 billion in online sales revenues in the United States and $3.2 trillion globally by 2003.

As one former Jupiter client says, "You use SPS if you like the figures they offer. Otherwise, you use other analysts' numbers."

For many, Jupiter is less a predictor of trends and more a validator of business plans. One industry observer puts it this way: "People pay money for [the reports] because it's better than nothing -- they need numbers, estimates of market size and market growth because business plans are based on this. It's in everyone's interest to show big markets and fast growth, so this stuff is skewed that way. People are buying market research because they need to show venture capitalists or board of directors that this is a hot market and that there's a lot of opportunity."

Jupiter's own S-1 filing includes a chart that demonstrates the growth of Internet commerce markets as proof of the size of its customer base, and quotes its own predictions about the growth of online travel.

Next Page: Do Jupiter's clients get preferential treatment in its research?

• View blog reactions