Boston Globe article. Julio is quoted about 3/4 of the way down.
Analysts Say Predicting Future of High-Tech Companies is Tricky
Feb. 21 (The Boston Globe/KRTBN)--One of the hottest Internet companies
of the past decade died quietly last week. PointCast, a company that
automatically "pushes" data to millions of Internet-linked computers,
had once been considered a world-changing company at the forefront of a
crucial technology.
Among those singing the praises of push technology were the technology
industry analysts at the Yankee Group in Boston. Customers who bought
the company's reports on push services would have learned that
PointCast and similar companies had a bright future. Indeed, in late
1996, Yankee analysts predicted that push services would be a $5.7
billion business by now. Instead the market quietly faded away.
On the other hand, in 1998, Yankee Group predicted that there would be
about 1.25 million households using high-speed Internet connections.
That forecastproved to be about right. According to recent cable
industry data, there are now approximately 1.4 million users of
high-speed cable modems in the United States.
Predicting the future is a perilous business, especially in
high-technology industries. But that hasn't stopped a host of expert
analysts from attempting it. Companies like the Yankee Group, Forrester
Research of Cambridge, International Data Corp. of Framingham, and
Jupiter Communications Inc. of New York earn millions for providing
businesses with forecasts on the high-tech future, and strategic advice
on how to prepare for it.
Nervous executives aren't the only ones who rely on these experts for
hire. Journalists from every newspaper in the land, including The
Boston Globe, routinely quote analysts from these leading research
firms, and build stories around the statistics and forecasts they
provide.
But some experts say that forecasts of industry trends are little more
valuable than horoscopes, especially when they attempt to look three or
four years into the future.
"There's no science in industry forecasting at all," says Boston
business consultant William Sherden, author of a skeptical book about
the folly of trying to predict the future. Even Sherden admits,
however, that he sometimes purchases a forecast, in search of a little
near-term guidance about industry trends.
Talk to people who produce the research, and they'll concede that
their long-term predictions are at best rough guides to the future.
John Gantz, chief research officer at IDC, even warns that in many
cases, nobody goes back to check the accuracy of an estimate made years
earlier.
"I don't care what anyone tells you," says Gantz. "No one in this
business goes back and looks at the actuals ... not on a rigorous
basis."
Still, Gantz insists that the very popularity of his firm's services
provide some evidence that IDC often gets it right.
"If we get it wrong often enough, they won't buy our stuff," he says.
Besides, Jupiter Communications chief executive Gene DeRose readily
acknowledges that the specific numbers aren't really the point.
"It's really about the lessons that we give them, and the dynamics of
how the industry works," he says. DeRose says his customers need the
assurance that statistics can bring.
"The marketplace demands and has a passionate need for these numbers,"
he says. But in the end, Jupiter and other firms say they use the data
mainly to point up trends.
Deborah Spence, assistant director of e-commerce marketing for Martha
Stewart Living Omnimedia Inc. in New York, is a loyal buyer of research
from Jupiter Communications, even though she knows their long-term
forecasts will sometimes fall short.
"Jupiter has been very responsive to us getting the information we
need," Spence says, "and it's all about timing."
But what good is timing if the data are wrong? Spence replies that she
doesn't expect market research data to tell her precisely what's going
to happen next year or the year after.
"I work with statistics and numbers enough to know you can pretty much
make statistics say whatever you want them to say."
Besides, she adds, each forecasting firm will inevitably come out with
a different statistic.
What matters to Spence is not whether the market for online furniture
sales goes up 30 percent or 40 percent, but whether it's likely to go
up at all. It's the direction she's looking for, not a particular data
point.
Sometimes, the forecasters don't even get the trends right. Yankee
Group wasn't alone in wildly overestimating public demand for push
software products like PointCast. Wired magazine produced a special
issue devoted to push software, claiming that it would transform the
way people used computers. Major software makers like Netscape
Communications Corp. and Microsoft Corp., desperate to catch the wave,
built push software into their Internet browsers. They all got it
wrong.
Still, the analysts keep trying, driven by intense demand from
companies desperate for a glimpse into the future.
"You have to have some way to plan," says Glantz. "No company can
operate without plans and strategies that go beyond a year."
IDC, Yankee, Jupiter, and many other forecasting shops follow a
business model called "syndicated research." Old-line consulting firms
like Arthur D. Little & Co. had long done in-depth research for their
large corporate clients. But starting in the 1970s, firms like IDC and
the Yankee Group hit on a way to provide the same service to smaller
clients. They'd prepare detailed reports on various industry segments,
then sell the information to anybody who cared to buy.
The Yankee Group, for instance, was founded in 1970 by Howard
Anderson, a Harvard Business School graduate who'd noticed that there
wasn't much solid data available on computer companies.
"At some point he determined that there was a lack of knowledge about
how technology was going to impact business," says current Yankee Group
chief executive Berge Ayvazian. So Anderson began cranking out reports
on his own.
Today, Yankee Group has about a thousand customers. Each subscribes to
one or more of the company's two dozen service groups. For instance,
there are analyst teams that specialize in media and entertainment
businesses, telecommunications in Latin America, and of course,
Internet computing. Each subscription sells for $25,000. That money not
only buys access to research, but also the researchers. Subscribers are
entitled to phone their analysts for detailed briefings.
This human contact is the real value of the subscription, according to
senior vice president Brian Adamik.
"We require our analysts to talk to our clients every month," he says.
IDC follows a similar business model, but with a much larger catalog
of 200 services. Even more than Yankee Group, IDC is associated with
statistics -- data on past performance of key technology markets, and
forecasts of how those businesses will perform years into the future.
IDC and other research firms get their information in much the same
way reporters do, with lots of legwork. Say you're trying to figure out
how many computers are being sold. IDC's John Gantz says his company
would approach the question from multiple angles.
"We survey users to find out 'What did you buy,' 'What do you have
installed today' ... and we even call the vendors to ask them, 'What
did you ship?'" Gantz says. "We have relationships with the big
distributors. We get information from them on how much did they move
[and] pricing."
But there's more. The analysts will also contact the companies that
make components that go into the devices. How many hard drives are
being made? For that matter, how many hard drive controller chips are
being made? By tracking the supply of the components used in a product,
companies like IDC hope to get an accurate picture of how many are
being produced.
The analysts tend to be people with MBA degrees and a knack for
qualitative analysis, or veterans of the industries they track. A
handful, like Jupiter Communications' DeRose, got their start as
journalists. Along with hard numbers, the analysts absorb anecdotal
data that give them a feel for the state of an industry. A certain
amount of this gut-level instinct winds up in these reports, and plays
a major part in the one-on-one discussions between analysts and their
customers.
Anecdotal data plays a large role in the reports issued by Forrester
Research. Much of the firm's research is based upon interviews with
vendors, distributors, and consumers, in an effort to understand their
buying patterns or marketing plans. The results are written up in
brief, breezy reports with relatively few hard numbers to slow the
reader down. (Forrester executives declined to be interviewed for this
story. The company is in a federally mandated "quiet period" because of
a pending stock offering.)
"It's like Harvard Business Review-Lite," says Jeffrey Rayport, an
associate professor at the Harvard Business School, and a consultant
with Monitor Group of Boston. "There are those who say that's totally
unscientific."
But Rayport is just as skeptical of the number crunchers like Yankee
Group and IDC.
"There are others who'd say the numbers you get ... are totally
suspect, so how can you trust those?" In Rayport's view, Forrester's
word-of-mouth approach is at least as likely to produce reliable
results.
Not all the technology analysts try to gaze far into the future. Julio
Gomez of Gomez Advisors Inc. in Lincoln prefers to sell customers
accurate information about the short term.
"We give them tools to help them do their jobs today," he says, "not
tell them what things are going to look like in five years."
Gomez's 40 researchers analyze the performance of a host of Internet
retailing companies, using a set of standardized criteria -- how easy
is it to use the site, for instance, and how its prices compare with
those of other sites. The resulting data is used to rate the quality of
these retail Web sites, and these ratings are constantly reviewed and
updated.
The final ratings are available to the public at the Gomez Web site.
But Gomez also uses the data to produce short-term analyses of trends
in various Internet retail markets, like online computer retailing or
bookselling. It's these reports that Gomez sells for a fee to corporate
clients.
"Our value is more for tactical decisions than for long-term strategic
decisions," Gomez says.
That might endear him to William Sherden, a renowned skeptic of
efforts to predict the long-term future. In his 1997 book "The Fortune
Sellers," Sherden mocks human efforts to make accurate predictions in a
wide range of areas, from weather to the stock market. In his view,
it's possible to make fairly reliable predictions about an industry's
performance looking ahead a few months, usually by assuming present
trends will continue. But go further out -- a year or more -- and it
all becomes guesswork.
Sherden isn't questioning the analysts' ethics.
"Ninety percent of the people who are producing these forecasts
actually believe they're doing something useful and valuable and
reliable," he says. He just thinks they're kidding themselves.
In any case, Sherden says, companies "are not buying it for market
predictions; they're buying it for technology trends."
It's an attitude that the technology forecasters cheerfully embrace.
"I agree that you can't possibly predict very accurately or with much
detail ... what that market opportunity or that new arena is going to
look like," says Yankee Group's Ayvazian. "It's not the forecast that's
valuable; it's the strategy that's valuable."
By Hiawatha Bray
-0-
To see more of The Boston Globe, or to subscribe to the newspaper, go
to boston.com
(c) 2000, The Boston Globe. Distributed by Knight Ridder/Tribune
Business News. MSO, AOL, MSFT, END!A2?GL-TECH-FUTURES |
