Mobile broadband: When is it profitable?
Mobile data plans have started to evolve. U.S. mobile operators, under pressure from exploding traffic volume in their networks and grappling with network congestion, have recently announced new service pricing for voice and data.
Though it is still unclear what the net effect of the new pricing will be on subscribers, the new plans set an important departure in the approach to voice versus data service plans. Subscribers may be able to get cheaper voice, but data pricing shows no comparable decrease.
More importantly, operators have started to require that subscribers have a data plan with some device types--mostly smartphones. Mobile operators seem to acknowledge that they can offer a better deal to their subscribers for voice services, which are substantially more profitable (with the exception of SMS, which on a per-bit basis may give an even better return than voice), but they have chosen not to extend this change to mobile data plans.
Why didn't data prices go down? Two reasons may explain the new approach to data. Mobile operators want data to become an integral part of subscribers' service plans. They are willing to charge less for voice, as long as subscribers pay on average more for data. This reflects a more balanced approach in splitting revenues between voice and data--with operators less willing to have voice subsidize data services. As data services become more mature and widely adopted, this is an approach that is no longer sustainable.
The second reason is that mobile operators cannot afford to lower data plans, as they may lead to a downward spiral in ARPU, at a time when they need to deal with an unprecedented growth in individual user traffic.
Initially the bulk of the growth in data traffic was generated by iPhone users, who now use more than 500 MB per month. With the introduction of new devices including Android-based smartphones and the higher number of applications, mobile data is rapidly becoming a mainstream consumer service. In the U.S., smartphones now represent over 30 percent of shipments and the percentage is likely to go further up. Traffic generated by these devices is also quickly catching up with that from iPhones. In Russia, mobile WiMAX operator Yota sees over 1 GB per month data traffic from subscribers using their HTC smartphone. For laptops, this figure is a staggering 13 GB per month. A large--and rapidly growing--portion of this traffic is video. This is what worries operators worldwide: with email and Internet access, traffic growth is bound by the limited requirements of the application; with video or even audio content takes little effort for subscribers to enter in the realm of the GB/month.
Mobile operators are delighted to see that their subscribers love mobile data services--and that are willing to pay for them--but at which point does the growing popularity of data and video affect profitability of the mobile data? It does not take long, as a quick back-of-the-envelope calculation that compared delivery costs and revenues on a per-MB basis.
The revenues per MB can be computed as a function of the monthly fees, for different levels of traffic. Revenues can be compared to the delivery cost per MB, which we estimate at $0.015 for HSPA, $0.005 for LTE and $0.003 for WiMAX on the basis of our analysis of mobile operator and vendor data. The delivery cost per MB depends on many variables that are specific to different operators--and in particular on the network utilization level--and therefore are subject to variability. Our values are therefore only indicative and on the low end in comparison with other estimates we have come across. It has to be kept in mind that these estimates do not include costs such as customer acquisition and support, or network core operations, which are shared with voice.
At the current average traffic levels of 500 MB/month, revenues per MB outstrip delivery costs for HSPA, LTE and WiMAX, for ARPUs starting at $20 per month. As traffic grows, however, the costs per MB rapidly exceed the revenues, especially when charging subscribers low fees. At a 50 percent CAGR, the 500 MB per month will reach 2.5 GB per month in five years. At $20 per month, for instance, mobile operators operate at a loss for subscribers using more than 1 GB per month in an HSPA network, or for subscribers using more than 5 GB per month in an LTE network. At 10 GB per month, data subscribers do not generate any net benefit for mobile operators with HSPA. With LTE or WiMAX, revenues from 10 GB subscribers at best reach the delivery cost...Continued
Comments
Great article. Are the major wireless incumbents reluctant to immplement QOS/traffic prioritization technology due to the resulting potential for VOIP cannibalization of voice services?
Where on earth does the cost per GB really come from? The cost per GB as reported or commented by operators today is merely a Cost to be allocated vs Measured traffic in more or less empty networks... How is that a relevant measurement??
As long as operators have bottlenecks outside of the HSPA Radio environment, there is no relevant cost to compare HSPA with LTE. The only relevance may be in comparing operators and where they happen to be on the "learning curve", i.e. how well utilized is their network?!
Agree to that, if the cost per MB was as high as it is often claimed by operators, how come eMobile in Japan are able to make a profit out of a business consisting almost entirely of PC-based users?
Where are the calculations to proove the claims on cost, as per this article or others? It's just hearsay and assumptions, no facts and no data...
Where does the number 50% CAGR in terms of traffic per user come from?? All previous reports with some real insight, including Ciscos global reports, say Traffic per Subscriber grows between 20 and 30% per Year. One can't mix traffic growth in the network with traffic per subscriber!!
At the same time, in fast growing HSPA networks (fast in terms of subscribers) the average traffic per Subscriber is going DOWN, not UP. Of course this will change again when subscriber uptake tapers off in a few years but until then none of this is true.
But perhaps the most important factor of all, you CAN'T make assumptions on Cost based on Traffic per subscriber. It's ONLY the traffic generated at Busy Hour that drives cost. The high average numbers in Russia for example are mainly generated by p2p traffic, and 23/24 roughly of that is outside of Busy Hour...
The growth rates are spreading the utilization of a minority of high usage users across the rest of the data customer base. With the launches of new services and devices on 3 & 4G networks, there will be an expected hockey-stick growth curve and will eventually subside somewhat...That is until the next new must-have product or device launches. As new apps are allowed to utilize wireless bandwidth such as video streaming service Slingbox, we will see this utilization go through the roof. If you build it, they will come as they say. Regarding the CISCO reports I cannot weigh in because I do not know the sample and scope of the data sets. I can say that the traffic growth is around 50% CAGR for wireless data if you spread total utilization across all data subs for one particular wireless service provider.
Wwe've been saying for a while over at www.broadbandexpert.com that data charges are too expensive, more in relation to mobile broadband plans, e.g. access via a usb mobile broadband plan, as opposed to mobile internet on your phone. There's a huge opportunity for the providers to monitise their 3G and 4G networks if they make mobile broadband plans affordable; in the US prices are 2 or 3 times high than most european coutires meaning that mobile broadband plans are really only taken out by business and are missing out on the huge consumer market.



SHARE
WITH:
Comments (6) | Post a comment