By Iain Gillott – Commentary
Much has been written and speculated in the last week about Sprint’s network plans after the publication of a Re/Code article and the plan to significantly cut network expenses. Aside from the resulting pressure on the stocks of several wireless companies, including Sprint’s, it also appears that the article has caused Sprint to move up its Q4 2015 earnings announcement. To summarize, the RE/Code article certainly put the wireless cat among the pigeons!
We have spent the last week talking with various wireless industry contacts to try and understand exactly what Sprint is planning. While Sprint may provide more detail on the earnings announcement tomorrow, we thought it worthwhile to tell you what we know now. The picture is not pretty and, to be frank, has led us to question the sanity of the plans. Sprint loves to name projects (‘ION’, ‘Network Vision’, etc.) so we would like to offer our own suggestion for the latest idea: ‘Network Suicide’.
Before getting to the plan details we believe are being discussed, I should note that, after talking to vendors, technologists and network planners for the past week, we spoke to no one (and I mean nobody) who believes this is a good idea. Obviously there are some who believe the plan is viable, no doubt including Sprint senior executives. Also note that iGR did ask both Sprint and Mobilitie for comment on the plans (we sent emails outlining the essence of what we believe to be happening on Friday January 22nd) – we did not hear back from Mobilitie. Sprint’s Corporate Communications department did reply saying that they expected many of the questions to be answered in the earnings call. We shall see.
Based on industry conversations in the past week, here is what iGR believes to be the details on Sprint’s new network plan:
– Significant macrocell site decommissioning, perhaps up to 80 percent of Sprint’s current macro sites. This seems extreme, but the 80 percent figure kept coming up in conversations.
– Replacement of the existing macro sites with so-called ‘mini-macros’. These are essentially macrocells mounted on 80, 100 or 120-foot metal poles.
– Some macrocells on the edge of the network (exactly where is unknown) will be decommissioned and not replaced since it appears the current traffic load does not warrant a cell in that location. Sprint appears to be relying on roaming partners in these areas for continued coverage.
– A high number of small cells, using existing or new wooden poles. These new poles will, again, vary in height. Existing telephone poles, light poles, etc., will be used where possible. This aspect of the plan is not really new and is not a unique idea to Sprint; the other operators have been looking hard at small cell locations and the use of poles for the last few years.
– By some accounts, a purchase order for the new poles may already have been issued.
– Much has been written in the past week about Sprint’s plans to put the poles and ‘mini-macros’ in the public right of way (RoW) and on government land. This aspect of the plan appears to be accurate. iGR believes that Sprint and/or its partners somehow plan to claim rights as a CLEC in order to use the RoW. Quite what the legal issues are here we do not know (lawyers, feel free to comment!) but iGR understands that Sprint is classed as a CMRS (Commercial Mobile Radio Services) provider like the other mobile operators. Quite how a CLEC designation can be claimed or used is unclear.
– Fiber backhaul will be replaced with NLOS microwave using Sprint’s own 2.5 GHz spectrum. While engineering plans are believed to be underway, it appears Sprint plans to aggregate the microwave links and then use dark fiber connections back into the rest of the network.
Why iGR believes the plan will not work
Sprint is trying to take a radical, even innovative, approach with the new plan, but many aspects of the approach have been tried before in the industry. That said, let’s discuss why, in iGR’s opinion, these ideas are unlikely to work. Remember Sprint’s goal is to cut the cost of operating its network as quickly as possible.
1. Tower leases are multi-year and termination costs need to be considered
The majority of Sprint’s existing macrocells are located on towers owned by a variety of companies; this model is used by the other mobile operators as well, obviously. Tower lease agreements are multi-year in length. The tower companies’ business model is basically that of real estate – build an asset and then lease access over many years to recoup the initial build cost. As many financial analysts have commented in the past week, many of Sprint’s tower lease agreements have three to five years left. If Sprint cancels these agreements, there will be significant penalties and cancellation fees that will incurred. iGR does not believe that the tower companies are likely to give Sprint any relief on the penalties simply because Sprint does not want the tower space anymore. Sprint signed the lease agreements (remember Network Vision reworked many of the macrocell configurations) and will, in all likelihood, be held to the terms of those agreements.
2. Ditto for the current backhaul agreements
As with the tower leases, Sprint also has multi-year agreements for fiber backhaul to the majority of the sites. Remember again that Network Vision reworked much of this backhaul. Now Sprint wants to use microwave backhaul? There will be cancellation fees and penalties associated with the current backhaul agreements.
3. Cost of commissioning and decommissioning
No one has discussed the decommissioning and installation costs. If the plan goes ahead, Sprint will need to pay someone to remove the existing equipment from the towers; that means climbing the tower, using a crane to remove the antenna arrays, removing the equipment at the base of the tower, etc. This will be expensive and the cost will be incurred in the short term.
Next, the new small cells and ‘mini macrocells’ have to be purchased installed, configured, tested and commissioned. Again, this cost will be incurred in the short term. In an ideal world, Sprint would do this before decommissioning the tower-based macrocells. Last year, iGR modeled the cost of deploying small cells on new and existing poles: the expense is not in the radio equipment but rather in the construction and deployment. This will be an expensive proposition for Sprint.
4. Poles in the Public RoW
For iGR, this is one of the most troubling aspects of the plan and the area with the most unknowns. If Sprint executives – here or in Japan – think this is a new approach, they are wrong. Many companies over the years have looked at using the public right of way (and any other RoW!) for locating cell sites. Railways, utilities, municipalities, commercial buildings…all have been explored at one time or another by various companies. Locating cells is difficult and takes time for a reason: the value of a cell site is now well known and few people want a tower or large pole behind their house!
So with that in mind, why do Sprint and Mobilitie believe that they can use the public RoW without facing problems? The scary proposition may be that, in the interest of time-to-market, Sprint does not ask for permission but simply puts up the new poles and then deals with the backlash later. This would be a huge problem for two reasons: poles and towers that are erected without permission can simply be removed (hence no network in that location) by the governing municipality. Also, many other companies are currently working within the system to get permission for other, similar projects. Sprint’s approach, if true, could poison the well for everyone and set the industry back years.
As many of you know, I live in Texas, where the major cities have already frowned upon plans that require more poles. As we were discussing Sprint’s plans with the industry this week, more than one person commented that Sprint’s pole-based network plan would never work in Dallas or Houston. Some people just started laughing!
The nightmare scenario for Sprint is that poles deployed in the public RoW without permission are then taken down by the respective authority or authorities. This would mean that Sprint had no coverage at that location (with the subsequent impact on customer experience) and would then need to locate a new cell. This would take time and money, both of which Sprint appears to be short of.
Can Sprint avoid upsetting any jurisdictions with this possible plan? Unlikely. Take New Jersey as a single example: there are 567 separate jurisdictions in the state. Could Sprint really put up thousands of new poles and hope no one notices or cares? Not on this planet in this lifetime.
5. Microwave backhaul is limited in capacity
It is ironic to iGR that Sprint is again discussing microwave backhaul. Clearwire built their network using microwave but they had the advantage of building a greenfield network and actually having no customers at the start. Sprint has a large existing network that needs to be replaced and 50+ million customers that expect service. Microwave backhaul will scale as traffic increases, but this takes additional spectrum (for the microwave radios). Sprint is believed to be planning to use its own 2.5 GHz spectrum; this will not require a microwave dish but instead use a small antenna. But, range is limited.
iGR believes that Sprint could engineer the microwave backhaul, in theory, to work. But what do they do in a few years when more capacity is needed, especially to support the increase in video traffic? How do they deal with the longer latencies of microwave when their competitors are moving to low-latency, high-bandwidth 5G? And what of Sprint Spark (in 18 cities) – will this be run on microwave backhaul? In short, iGR believes the microwave backhaul strategy essentially exchanges a potential fingers-crossed short term gain for long term network growth and viability. It is short sighted.
6. The location of a cell site matters
Cell sites are located where they are for a reason: that is where the people are and hence where the network demand is. You cannot simply remove one or two towers and move the cells without impacting the network performance and customer experience. If Sprint’s plan were viable, would you not think other operators over the last 10 years would have tried something similar? Perhaps there is a reason they have not done this. Cell towers cannot be placed where you WANT them to be – they are located where they NEED to be.
7. The U.S. and Japan are very different
Sprint executives have been very public about their goal of bringing Softbank’s Japanese network practices to the U.S. and specifically the use of small cells. While this sounds good in practice and no doubt there is something to be learned, if the current plan is anything to go by then Sprint’s executives are not understanding the situation.
The U.S.’ geography is very different from Japan. Japan is relatively small (slightly smaller than California). Moreover, 90 percent of Japan’s population lives in cities with a relatively high population density. About 80 percent of the U.S. population lives in cities, but the U.S. population density is less than half of Japan’s, even in the cities. Consider that the population density in Tokyo is more than double that of New York City, which has the highest density in the U.S.
So, taking a similar network approach as used in Tokyo and using it in the U.S. is unlikely to be successful. RF networks are hard to engineer and RF signals are particular; physics dictates how they behave and physics does not always cooperate. Mobile operators engineer their networks differently across the U.S. to account for variances in climate and geography. If Sprint’s executives believe they can replace macro towers with small cells and ‘mini macros’ just because it worked in Japan, they are likely in for a rude surprise.
For an example of how to quickly improve a network’s performance in the U.S., Sprint execs should be looking at T-Mobile USA. TMUS took some spectrum and cash from the failed marriage to AT&T and leveraged it extremely well into an LTE network that continues to win awards. This has included support for a video service. And this was done through the use of existing towers and improved fiber backhaul, among other moves; T-Mo is not yet known for its use of small cells. The result has been, of course, that T-Mobile has taken Sprint’s spot as the #3 U.S. provider.
8. Roaming costs are likely to increase
One irony of this plan appears to be that some macro cells (an unknown number) at the edge of the network will be decommissioned and Sprint will rely on roaming partners. This means roaming costs will increase. If the new poled-based cells are then removed by force from public RoWs, roaming costs could increase still further. Future population and metro-area growth could make this problem worse. What is on the edge of the network today could be part of the metro market tomorrow (look at north Dallas for an example).
9. Competition is intense
One of Sprint’s major problems is that time is not on their side. iGR does not believe there is time for another major network rip-and-replace strategy. Unlike in years past when Sprint had to merge (with Nextel) or improve the network (Network Vision and Sprint Spark), the mobile operator competition is organized, focused and effective – more than it has ever been.
AT&T, Verizon Wireless and T-Mobile have all laid out defined strategies and are executing on them (obviously each have different approaches). Each carrier offers strong financial incentives to consumers who want to churn. Each offers variations on mobile video and are targeting consumers like my kids (18 and 21 years old) who consume mobile video as easily as breathing. (My son used 18 GB of LTE data over the holiday in TWO WEEKS).
Sprint and Softbank executives have stated that they believe the U.S. networks are too congested to support effective mobile video delivery, but iGR (and my son) will say they are wrong. Networks certainly have their issues, but consumers can effectively stream video over LTE – and they are doing it today.
Even if Sprint’s new rip-and-replace network plan works, it will result in years of network upheaval and uncertainty for consumers. In the meantime, the other three nationwide mobile operators, and many smaller ones, will be eating Sprint’s lunch. Remember that the other operators are expected to participate in the 600 MHz auctions and they already have AWS-3 spectrum to action. Sprint simply has no time for a massive new network revamp.
10. Churn risk is significant
Finally, lets talk about the consumer. As you may know, iGR conducts a lot of end-user research and recently surveyed U.S. consumers as to why they churned and, importantly, what will cause them to churn. Second only to price is a change in the network experience. Note that a change in a consumer’s network experience, not simply a worsening network experience, can cause significant churn. So if, as a result of this network plan, a consumer drops calls in a new location while maintaining calls in a location where previously they dropped, this will be seen as a net negative and will likely cause churn. Consumers want network improvement across the board, not parity.
The problem, of course, is that Sprint risks significant changes to network experience while this plan is being executed. Coverage could easily worsen before it improves. And Sprint’s consumers are unlikely to be willing to accept a ‘Be patient, we are trying, things will improve’ message from Sprint; competition is too strong for that to work. Any network plan that Sprint enacts must improve the network without significant risk of short-term degradation. This plan is not it.
Wrapping it all up
As you may have worked out, we are not fans of Sprint’s current plans as they appear to be. It is not clear how this plan in any way results in reduced network operations costs for Sprint. In fact, we believe that costs will increase in the short term if this plan is realized.
To be frank, we hope for Sprint’s sake that we are wrong in the analysis presented here. iGR believes that the significant decommissioning of macrocells is unlikely to reduce costs and much more likely to significantly degrade the perceived quality of Sprint’s network.
But, if all the people we have spoken to in the last week are to be believed, then plans we have outlined are currently accurate. If enacted, this plan paints a very scary future indeed for Sprint.
About the Author: Iain Gillott, founder and president of iGR, is an acknowledged wireless and mobile industry authority and an accomplished presenter. Gillott has been involved in the wireless industry, as both a vendor and analyst, for more than 20 years. The company was founded in 2000 as iGillottResearch in order to provide in-depth market analysis and data focused exclusively on the wireless and mobile industry. Before founding iGR, Gillott was a group VP in IDC’s telecommunications practice, managing IDC’s worldwide research on wireless and mobile communications and Internet access, telecom brands, residential and small business telecommunications and telecom billing services. Prior to joining IDC, Gillott was in various technical roles and a proposal manager at EDS (now Hewlett-Packard), responsible for preparing new business proposals to wireless and mobile operators.