How important is accurate telecom forecasting?
Sitting through another presentation this week on the explosive growth of mobile data, and the eye-popping statistics about how many Facebook and Twitter users there are, I can’t help but be excited by the possibilities. But I also have this feeling of déjà vu and a little voice in my head repeating the name Andrew Odlyzko.
If you were involved in the telecom industry about a decade ago you may be having the same feelings. From the mid 1990s until 2000 the Internet was growing by leaps and bounds. I must have sat through 100 PowerPoint presentations that repeated the same statistic: Traffic on the Internet is doubling every 100 days. I never heard anyone question that data. Seasoned telecom executives believed it, government officials believed it, the mainstream press believed it. Major corporations bet the farm based on that number. We had all become web and e-mail junkies, so it must have been true. Andrew Odlyzko was a mathematician at AT&T Research Labs. He was studying actual data usage within the core of the Internet. His conclusion was that Internet traffic was at best doubling every year. Doubling every year is great growth for a market, but if your forecast is for 800 percent sooner or later you are going to be severely disappointed. Odlyzko’s cautionary findings were published in several journals, but who wants to talk about tomorrow’s hang-over when the party is in full swing.
And what a hang-over it was, over a trillion dollars disappeared from the market cap of telecom firms and their suppliers and over 500,000 employees were given pink slips. In some places the excess fiber is still not being used.
The point of this trip down memory lane is that when something new is growing very fast, it can be hard to tell just how fast. If you are responsible for directing the investments in infrastructure to meet a growing demand then it is very important, not just to estimate the total market, but to break forecasts all the way down to lower levels of infrastructure, devices, and personnel.
We all know that the demand for mobile broadband is growing very fast. But how fast? Cisco publishes a Mobile Data Traffic Forecast, and their latest data published on February 1 claims a 2010 growth rate of 159 percent. From now until 2015 Cisco projects the Cumulative Annual Growth Rate (CAGR) to be 92 percent. But what if it turns out to be only 72 percent or maybe it grows 112 percent. If you are planning upgrades that take 9 months to execute, you can wind up with either unused infrastructure, or worse: dissatisfied customers.
This is why a strong forecasting system, using real data is so important. It is easy to get excited by the hype and believe growth projections, because we spend most of our time around people who love their mobile devices. But don’t forget, bubbles can burst. It is a lot better to be the one who had good forecast information.