"Steady increase".
Makes a welcome change from "exponential" and similar hyperbole in the past.
Cisco has put out its revised 2012-2017 forecasts for mobile data traffic. I'm horribly busy today so I can't put out a forensic analysis or listen in live to the webcast.
However, a couple of headlines:
Cisco has reported global mobile data growth of 70% during 2012, and revised down its 2013 mobile data traffic forecasts from 2.4EB to 1.6EB, and 2016 forecasts from 10.8EB to 7.4EB (EB=Exabyte). I think it's still being over-enthusiastic, but this is a welcome return to a sense of reality.
"Tiered plans are effective" - to me, this is a key point of understanding and change to Cisco's methodology. Basically, caps and tiers are effective at changing subscriber behaviour. More fancy approaches are (largely) unnecessary - and as I've often written before, this indicates that all the pain associated with unworkable application-specific tariffs can be avoided.
Also, it has recognised that WiFi use is much higher than first thought - although it's a little unclear whether it is just counting "true offload" (ie data that would otherwise have gone over cellular), or also includes what I term "private WiFi use" and extra elasticity, as users exploit free/fast connections by using more data than they otherwise would have.
Where I have doubts is about its forecasts of usage per device, which don't appear to have factored in late-adopters (& emerging market users') likely much lower usage than more "experienced" or "enthusiastic" smartphone owners.
While I applaud Cisco for (perhaps belatedly) down-grading its forecasts, I'm not convinced that deleting its earlier numbers and predictions from its website was such a wise move. As a forecaster, you need to be able to be proud when you get it right, and humble when you get it wrong. (FAQs: "Historical IP traffic numbers attributed to the Cisco VNI are listed in Wikipedia" - why?) I found a mirrored version of last year's full report here.
Overall, I'm feeling pretty vindicated. I put out a long blog post last year on why I thought many forecasts were over-cooked, and it's something I've gone into greater depth with my consulting clients. It's good to see some revisions - now we've got past the "tsunami" hype we can focus on how to make the right investment / network / service packaging decisions.
Edit: great & detailed analysis by Tim Farrar here
(Sidenote: Qualcomm, now *please* stop with the 1000x marketing slogan. You're now at the top of the forecast-inflation league table)
Makes a welcome change from "exponential" and similar hyperbole in the past.
Cisco has put out its revised 2012-2017 forecasts for mobile data traffic. I'm horribly busy today so I can't put out a forensic analysis or listen in live to the webcast.
However, a couple of headlines:
Cisco has reported global mobile data growth of 70% during 2012, and revised down its 2013 mobile data traffic forecasts from 2.4EB to 1.6EB, and 2016 forecasts from 10.8EB to 7.4EB (EB=Exabyte). I think it's still being over-enthusiastic, but this is a welcome return to a sense of reality.
"Tiered plans are effective" - to me, this is a key point of understanding and change to Cisco's methodology. Basically, caps and tiers are effective at changing subscriber behaviour. More fancy approaches are (largely) unnecessary - and as I've often written before, this indicates that all the pain associated with unworkable application-specific tariffs can be avoided.
Also, it has recognised that WiFi use is much higher than first thought - although it's a little unclear whether it is just counting "true offload" (ie data that would otherwise have gone over cellular), or also includes what I term "private WiFi use" and extra elasticity, as users exploit free/fast connections by using more data than they otherwise would have.
Where I have doubts is about its forecasts of usage per device, which don't appear to have factored in late-adopters (& emerging market users') likely much lower usage than more "experienced" or "enthusiastic" smartphone owners.
While I applaud Cisco for (perhaps belatedly) down-grading its forecasts, I'm not convinced that deleting its earlier numbers and predictions from its website was such a wise move. As a forecaster, you need to be able to be proud when you get it right, and humble when you get it wrong. (FAQs: "Historical IP traffic numbers attributed to the Cisco VNI are listed in Wikipedia" - why?) I found a mirrored version of last year's full report here.
Overall, I'm feeling pretty vindicated. I put out a long blog post last year on why I thought many forecasts were over-cooked, and it's something I've gone into greater depth with my consulting clients. It's good to see some revisions - now we've got past the "tsunami" hype we can focus on how to make the right investment / network / service packaging decisions.
Edit: great & detailed analysis by Tim Farrar here
(Sidenote: Qualcomm, now *please* stop with the 1000x marketing slogan. You're now at the top of the forecast-inflation league table)
0 comments:
Post a Comment