Besides operators taking conscious steps to replace DSL with LTE, savvy customers are also jumping on the wagon realizing that LTE may provide an equally stable and fast experience with the added benefit of mobility. The most widely covered case is Japan, where NTT was forced to cut prices of its fiber-to-the-home (FTTH) service by 34% to stem the flow of its FTTH customers moving to LTE only (reportedly mostly youth)4. This evidence from Japan does not foretell the death of fixed broadband (far from it), but it does show that LTE has the ability to make a significant impression on the broadband marketplace when it is available on a national scale.
Key players in the US market seem to think so: together with Verizon discontinuing naked DSL while planning to have LTE deployed across the entirety of the 3G network by 2014, AT&T’s CEO Randall Stephenson admitted that DSL is an “obsolete technology, which is slowly being replaced by wireless and fibre”5. In November 2012 AT&T has decided to drop DSL, as the company believes that an all-IP network is the path to more profitable future.
But the US market is fairly peculiar: cable is the incumbent in most areas, even rural (cable has ~40% market share in the broadband residential market), and offers faster speeds than the legacy DSL of Verizon and AT&T. In these areas LTE is a viable alternative, given the technology’s superiority and its capability to provide high enough speeds to serve the Home and synergies with Verizon’s and AT&T’s mobile businesses.
Beyond the USA (and Germany, though mostly regulatory driven) cases of LTE substituting DSL for last mile connectivity are still few and far between. One of the few worth noting is Australia, where Telstra is complementing its newly launched NBN’s fibre network with LTE as an alternative to ADSL+ to serve the traditional fixed broadband market. To do so, it is extending its LTE footprint to two-thirds of Australia’s population by the end of 2014, with the objective to capture a segment of the traditional fixed broadband market.
AT&T, Verizon and Telstra’s thinking is understandable: LTE provides advantages over DSL and can be a viable alternative even to FTTH, as long as some conditions are met. We analyse them in the following section.
The killing zone for DSL
1. Low density - rural deployments
Theoretically LTE can provide speeds similar to DSL but they can be variable and difficult to guarantee given the customer mobility and the fact that bandwidth is shared amongst customers in the same area (see exhibit 1). Other elements such as topography, buildings, indoor vs. outdoor coverage and weather can also degrade LTE speeds (not factored in the next graph).
On the other hand, DSL speeds are usually dependent on the quality of the copper and backbone contention, together with the distance from the exchange, both of which can be predicted.
This means that in urban areas, where household density is high and fixed broadband is generally available LTE is likely to be a sub-optimal solution for home and SME connectivity.
However, in rural areas, where population density is limited (and hence the number of users per cell is low enough), LTE deployed on a lower spectrum is likely to deliver a better service than DSL. Hard-to-reach rural DSL subscribers generally get much slower speeds than city dwellers, given their high distances to local exchanges and delayed technology upgrade cycles reducing the speed of the service. In these cases LTE can assure good enough average speeds to guarantee a satisfactory customer experience, allowing users to enjoy most services, including streaming Video-On-Demand (VOD).
In Germany, where LTE deployments have so far been focused on less dense areas, LTE shows average speeds of around 10Mbps6, proving to be a valid alternative for a typically sluggish rural DSL service (between 1 and 6 Mbps).
Besides providing better speeds, LTE can be very efficient in rural areas from a cost perspective given its wider coverage. LTE (and wireless technologies in general) has low incremental costs per connection besides the CPE (Customer Premises Equipment) and some variable cost in international connectivity.
On the other hand, adding a DSL subscriber implies relevant CAPEX, as it requires laying a copper wire from the nearest cabinet (or the exchange) to the customer’s premises. Additionally, DSL comes with high maintenance costs (OPEX), which make it economically unattractive to deploy in areas with a low number of subscribers and no business customers, i.e. rural areas.
As density increases, though, the balance shifts towards fixed (FTTC) due to capacity constraints: to provide the same speed on LTE cell density needs to increase and this negatively impacts the business case (see exhibit 2).
2. Availability of low spectrum requency
Spectrum plays an important role in determining the benefits of LTE vs. DSL, both from a QoS and financial point of view. While LTE deployed on lower spectrum (e.g. 700 / 800 / 900 MHz) can deliver speeds similar to DSL in sparsely populated areas, higher spectrum deployments (generally more suitable for capacity purposes than coverage) will generally deliver poor indoor QoS. In the latter case, players can still mount dedicated LTE antennas on the customer premises’ roof to capture the 4G signal (as Verizon does in the USA) or increase the number of LTE base stations, but either solution implies higher costs. In fact, the majority of the examples of rural LTE deployments mentioned beforehand (e.g. Verizon, Vodafone Germany, Telstra) leverage on low frequency spectrum bands.
The problem is that lower frequencies are generally already utilised. The 900 MHz band is used for GSM in the majority of countries, and only very few countries use it for LTE – Australia’s Telstra, Japan’s Softbank and Norway’s Tele2 and Telenor are some examples. The 700 and 800 MHz bands are the Digital Dividend bands, and so far only a few countries have refarmed and made them available for LTE – less than 20 players worldwide have commercial deployments or are testing these two bands7 (e.g. Verizon, AT&T and Sprint in the USA, TIM and Vodafone in Italy, Vodafone, DT and O2 in Germany, Optimus in Portugal, etc. while the majority of countries do not plan to action the Digital Dividend spectrum before 2015).
As the Digital Dividend spectrum becomes available, we expect to see LTE increasingly taking over in low density areas.
3. Greenfield deployments / Too costly DSL infrastructure
The cost argument mentioned in Point 1 applies to greenfield deployments and to legacy network is too costly to maintain.
In greenfield deployments operators would never consider deploying DSL at this stage, as Fibre and LTE are more modern and cost efficient technologies. However areas where fixed infrastructure is unavailable are rare in developed countries. Some studies place the number of European households not covered by DSL at 5%8 and the average number of households that have to rely on connections offering less than 5 Mbps at 20%9. In developing countries though fixed infrastructure can be more limited, making the potential for disruption is bigger (for example, Indonesia registers only 3% of the households covered by DSL, Egypt 8%, Mexico 40%). In fact, many emerging markets have been using WiMAX, CDMA or proprietary broadband wireless technologies for years, in order to cover the gap generated by limited fixed infrastructure. As we saw earlier, LTE can provide much higher speeds at lower costs than current WiMAX technologies, making it a game changer in these situations.
For areas where DSL infrastructure is in place, moving to LTE can be a more cost-efficient solution. Delta Partners’ own experience shows that the trade-off between shedding DSL clients and having to sustain the cost of running a DSL network often plays in favour of LTE. In fact, keeping DSL as a technology solution on top of fibre (which is required in urban areas and for corporate clients), and possibly wireless solutions (if the operator is integrated) increases the complexity of the technology mix, and requires a dedicated team to run and maintain it. If the number of subscribers on DSL is not high enough, LTE can easily become a more profitable option.
4. Low usage customers
LTE packages have affordable entry prices and, for comparable speeds, an LTE plan (dongle/router) carries a minimal premium (less than 20%) over DSL or is occasionally cheaper (see exhibit 3).
However, a real price comparison between DSL packages and LTE is unfair. Contrary to fixed broadband, LTE products tend to be volume-based, having caps often below 10 GB, which is less than the average household consumption. For instance, in March 2011, AT&T reported that its DSL customers used an average of about 18GB per month, and more recently 21GB10, while Cisco forecasts worldwide broadband consumption to more than triple by 2015, reaching 69 GB11 (see exhibit 4).
Operators cap LTE products on volume to avoid fixed-mobile substitution, to control quality of service and assure profitability remains above the desired level. This means LTE is a viable alternative to DSL for light users only – those who need connectivity to check emails, do social networking and some browsing. In fact, NTT’s subscribers who have abandoned fibre-to-the-home seem more prone to watching short-form videos than streaming or downloading long-form videos^12.
As such, LTE as a replacement for fixed connectivity is more suitable for countries where average broadband consumption is lower (for instance, 9 GB in the Philippines, 15 GB in East Africa) or for less sophisticated segments.
LTE has become the fastest growing wireless technology so far and has started a wave of fixed-mobile substitution, but its effect may not be so radical nor so imminent, due to a number of factors, including: 1) the intrinsic limitations of wireless technologies, 2) the lack of availability of low frequency spectrum to deploy LTE effectively, 3) the wide-spread presence of competitive DSL infrastructure in mature markets, and 4) the need for usage caps.
However, LTE still remains a viable alternative in a number of cases:
•Mobile-only players that want to tap into the Home market (both in rural and urban areas): LTE offers a new potential revenue stream to mobile-only operators without fixed infrastructure. Mobily is a fitting example. Contrary to the other two players in the market, it started its LTE deployment in smaller cities and offering “millions of [remote] home users across Saudi Arabia access to fast broadband speeds that will exceed speeds of a fixed line connection”13
•Mobile-only players already serving the Home market through partnerships with fixed line players, but who want to cut the cord with their fixed partners for last mile connectivity. LTE is likely to be.
cheaper than leasing DSL from the fixed incumbent – see the case of Vodafone Germany where its CEO Friedrich Joussen believes LTE is a much better investment than DSL as it avoids paying 10 Euro per month to DT for every last mile fixed-line connection leased
•Fixed providers wanting to expand into rural areas currently not covered by fixed technology (white spots) – Fitting case is Verizon’s program called “Connecting Rural America”, aimed at providing home connectivity via LTE to the 19 million rural Americans who don’t have Internet access at home^14
• Fixed providers wanting to boost their residential market penetration by offering lower-end urban /sub-urban areas an affordable broadband solution – Tele2 in Sweden is the perfect example where the operator has a small share of the DSL market (~10%), and decided to launch LTE for the Home as a way to elbow out DSL rivals, providing greater speeds at more affordable pricing