Comments on: Still going down? https://tmfassociates.com/blog/2014/01/30/still-going-down/ Satellites, spectrum and other stuff Tue, 03 Feb 2026 21:36:32 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: daveburstein https://tmfassociates.com/blog/2014/01/30/still-going-down/comment-page-1/#comment-11075 daveburstein Tue, 04 Feb 2014 10:55:09 +0000 http://tmfassociates.com/blog/?p=4815#comment-11075 Tim Excellent work, again Dave Burstein Tim

Excellent work, again
Dave Burstein

]]> By: timfarrar https://tmfassociates.com/blog/2014/01/30/still-going-down/comment-page-1/#comment-10711 timfarrar Fri, 31 Jan 2014 23:46:24 +0000 http://tmfassociates.com/blog/?p=4815#comment-10711 Yes, absolute growth is certainly increasing and it would be surprising if the increase in traffic wasn't larger each year than the year before. My expectation is that you have an S-curve (growing smartphone penetration) overlaid on an ongoing moderate annual usage growth per person (30% would be inline with fixed network experience, but it might be lower because much of the extra usage is carried on WiFi). My post is really directly to the accuracy of Cisco's predictions for a particular year (i.e. 2013). Their expectations for traffic levels in 2013 have fallen in each of the last two years, in contrast to the increases in earlier forecasts. Analyst forecasts often undershoot when growth is accelerating and then overshoot when growth is decelerating, and Cisco is no different. I'm told that Cisco's 2012 numbers have been retrospectively downgraded as well, because the expectations for laptop data traffic were overstated. Once these figures are available it will be possible to plot how the expectations for growth in 2013 (in addition to the absolute growth shown above) have varied over 6 years of Cisco VNI reports. Yes, absolute growth is certainly increasing and it would be surprising if the increase in traffic wasn’t larger each year than the year before. My expectation is that you have an S-curve (growing smartphone penetration) overlaid on an ongoing moderate annual usage growth per person (30% would be inline with fixed network experience, but it might be lower because much of the extra usage is carried on WiFi).

My post is really directly to the accuracy of Cisco’s predictions for a particular year (i.e. 2013). Their expectations for traffic levels in 2013 have fallen in each of the last two years, in contrast to the increases in earlier forecasts. Analyst forecasts often undershoot when growth is accelerating and then overshoot when growth is decelerating, and Cisco is no different.

I’m told that Cisco’s 2012 numbers have been retrospectively downgraded as well, because the expectations for laptop data traffic were overstated. Once these figures are available it will be possible to plot how the expectations for growth in 2013 (in addition to the absolute growth shown above) have varied over 6 years of Cisco VNI reports.

]]> By: Stephen https://tmfassociates.com/blog/2014/01/30/still-going-down/comment-page-1/#comment-10710 Stephen Fri, 31 Jan 2014 23:29:14 +0000 http://tmfassociates.com/blog/?p=4815#comment-10710 Tim, Good and timely observations, as always! Thank you! However, absolute traffic volume can increase while the percentage declines. The denominator in the growth rate is accumulating a large base. Only with exponential growth does the rate stay constant, the Cisco data has approximated a parabola, which is increasing traffic but less than an exponential function. For example, In their 2012 report, they show global mobile data and internet traffic going from: 2012 0.885 EB/mo 2013 1.578 EB/mo an increase of 0.693 EB/mo or 78.3% over 2012 # 2014 2.798 EB/mo an increase of 1.22 EB/mo or 77.3% over 2013 #. That's a slight decrease in growth rate but nearly double the absolute growth (1.22/0.693 =1.8). Using their new 1.5 EB/mo (max) number for 2013 we see an increase in traffic of (1.5-0.885=0.615 EB/mo) which is still over twice the increase seen in going from 2011 to 2012 (0.885-0.597), using their May 2012 report for 2011 traffic. I agree that growth rates are decreasing but the absolute yearly increase in traffic is still growing impressively. The equipment that is tied proportionally to capacity should be growing in unit sales year over year. However, price errosion is constraining infrastructure spending. Steve Tim,
Good and timely observations, as always! Thank you!

However, absolute traffic volume can increase while the percentage declines. The denominator in the growth rate is accumulating a large base. Only with exponential growth does the rate stay constant, the Cisco data has approximated a parabola, which is increasing traffic but less than an exponential function.

For example, In their 2012 report, they show global mobile data and internet traffic going from:
2012 0.885 EB/mo
2013 1.578 EB/mo an increase of 0.693 EB/mo or 78.3% over 2012 #
2014 2.798 EB/mo an increase of 1.22 EB/mo or 77.3% over 2013 #.
That’s a slight decrease in growth rate but nearly double the absolute growth (1.22/0.693 =1.8).
Using their new 1.5 EB/mo (max) number for 2013 we see an increase in traffic of (1.5-0.885=0.615 EB/mo) which is still over twice the increase seen in going from 2011 to 2012 (0.885-0.597), using their May 2012 report for 2011 traffic.

I agree that growth rates are decreasing but the absolute yearly increase in traffic is still growing impressively. The equipment that is tied proportionally to capacity should be growing in unit sales year over year. However, price errosion is constraining infrastructure spending.

Steve

]]>