MARKET RESEARCHERS ERRED IN PREDICTIONS ON DSL VS. CABLE MODEMS
When broadband deployment began to accelerate in late 1990s, market research companies flooded market with predictions promising dramatic growth for DSL but cooling in popularity of cable modem offerings. Some analysts forecast number of DSL subscribers would be twice as high as it actually is today. Others thought DSL would gain lead over cable by 2001, although cable modem subscribers still outnumber DSL subscribers by 2 to 1. In our examination of accuracy of market research studies in period (CD Aug 14 p4, Aug 13 p1), we encountered many explanations from analysts as to where their predictions went wrong but found little acknowledgment of any fundamental problems. One analyst claimed 34% error was “slight,” while another said no one could be expected to accurately predict deployment rate of new technology, even though his company and others sold their projections for up to $3,500 and more and rely on them to generate consulting relationships.
Sign up for a free preview to unlock the rest of this article
Timely, relevant coverage of court proceedings and agency rulings involving tariffs, classification, valuation, origin and antidumping and countervailing duties. Each day, Trade Law Daily subscribers receive a daily headline email, in-depth PDF edition and access to all relevant documents via our trade law source document library and website.
Jupiter was willing to acknowledge it was too bullish in its 1998 study predicting number of U.S. DSL households would hit 2.2 million in 2000. Best number we could find suggested U.S. had only 1.7 million households subscribed to DSL by year-end 2000, 23% less than Jupiter’s prediction. Dylan Brooks, Jupiter’s broadband analyst, said it based forecast on adoption rate of related technologies “such as cable modems, VCRs and PCs.” Brooks said DSL didn’t reach numbers of those other technologies because of “the lack of a killer broadband application,” even though that limitation also could be applied to cable modem deployment, which actually grew faster than most analysts’ predictions. When calculating actual number, however, Brooks said Jupiter extended current growth rate into future. DSL had been growing at a 40% rate every quarter and no one at Jupiter anticipated that rate would drop to 20%, he said.
Locking in growth rate is common error made by far too many analysts, said James Alleman, prof.-dir. of research at Columbia U. “These guys guess a lot and don’t do any analysis,” he said: “They just take an accelerating trend and make a forecast out of it.” He questioned whether many analysts were grounded in basic research and statistical techniques.
Analysts universally agreed that forecasts should not play major role in corporate business plans. Strategis Group analyst Adam Guy said using report required balance, for if businesses “don’t consider market predictions, they can miss the boat; if they rely too much, they can bet and burn.” He admitted it was “easy” for analysts to promise big numbers several years ahead.
Forecast accuracy generally was higher in short term, although at times even few months in future proved too challenging. In Feb. 2000, TeleChoice projected expansion of U.S. DSL lines in service to 2.1 million by year-end, 10-month projection. TeleChoice numbers in May 2001 showed there were 2.3 million DSL lines at end of 2000, suggesting fairly accurate forecast. However, another report by company in Feb. 2001 projected 5.7 million DSL lines by end of year. Despite again having advantage of short 10-month projection window, TeleChoice DSL Analyst Adam Guglielmo admitted actual numbers this time were likely to fall far short of that. “We didn’t see the economy declining this much,” he said, mere 6 months after report’s release: “We were thinking that the decline would slow down this year.” Guglielmo said: “We are a strategic consultancy firm, not a commercial organization that is trying to sell its reports.” Instead, he said, TeleChoice works with service providers “who want to have a clear picture” to plan their businesses.
Sometimes research firm projected numbers and sold them even while saying that such prognostications were impossible. In May 1999 Cahners study projected 2 million new DSL subscribers would be signed up worldwide in 2000. In fact, prediction fell far short, as there were 3.45 million new subscribers that year. Cahners Senior Analyst Ernest Bergstrom said that in 1999, as DSL was just beginning to deploy, nobody could have predicted how “new” and “challenging” market would unfold.
Consistent error among analysts was predicting DSL would quickly close gap on cable modem service. Cahners predicted DSL subscribers would surpass cable subscribers worldwide in 2001, saying: “Signs are beginning to emerge that DSL is ready to overtake cable modems.” However, FCC report on advanced services released last week (CD Aug 10 p4) said that while U.S. DSL lines increased to 2 million by the end of 2000, cable modem connections had risen to 3.6 million. Now, when there still are twice as many cable modem households as DSL homes, Cahners’ Bergstrom has made new prediction: That DSL isn’t going to overtake cable until middle of 2002 worldwide and end of 2002 in U.S.
Analysts who underestimated cable modem growth didn’t always acknowledge prediction’s shortfalls. In 1998, Jupiter report predicted number of cable modem households would reach 2.9 million in 2000, but according to TeleChoice that figure actually reached 3.9 million, surpassing Jupiter’s prediction by 34%. “Cable exceeded our expectations,” Jupiter broadband analyst Brooks admitted. However, he insisted he was satisfied with prediction: “We slightly underpredicted the broadband growth… just by one million.”
There were few reports that predicted cable’s continued dominance. Yankee Group in 2000 expected cable still would hold 58% of the market by 2004. Oct. 2000 Forrester report said that while cable would lose 30% of its market share by 2005, it still would lead market with 48% share, at 22.42 million households, with DSL’s share reaching 38% in same year. Guglielmo said he was “not sure if DSL is going to overtake cable at all. If it does, I am not sure when it will happen.” -- Susan Polyakova
Editor’s Note: This is edited version of 3rd of 3-part series in Communications Daily’s affiliated Warren’s Washington Internet Daily on market analysts and their predictions. On Mon. we summarized Part 1 on wireless data, voice-over-IP and domain names, and on Tues. we summarized Part 2, on broadband market.