Lowenstein's View: Can wireless compete in broadband?

Mark Lowenstein

SoftBank CEO Masayashi Son, in a Washington tour in March, implied that one of his key goals for Sprint is to become a competitive alternative in broadband, not just wireless. DISH's efforts to enter the mobile space are also predicated on using wireless as a way of offering some form of home broadband service--a key missing link in its residential offerings.

With expansion of LTE coverage and capacity, carrier aggregation, and new antenna technologies such as intelligent MIMO antennas, the technologies are coming together such that for the first time, we can actually have a serious discussion about mobile becoming a competitive alternative to fixed broadband.

Is it possible that a few years from now, we won't need to have separate subscriptions for fixed and mobile broadband service? Appealing as this prospect might be, my current view is that mobile could offer a viable broadband alternative for a certain segment of the market, but not for mainstream broadband households. But it's worth some examination from the standpoint of the market, technology, and economics.

The Market Situation

From a market standpoint, greater competition in broadband would be a good thing. The FCC and DOJ love to focus on how concentrated the U.S. wireless industry has become, and are likely to block a Sprint/T-Mobile merger, if proposed. However, wireless is way more competitive than broadband. The average U.S. consumer has a choice of four national facilities-based wireless providers, plus some regional players and scores of MVNOs. In broadband, most households are served by two ISPs, although in at least half of those situations, one of the operators' offerings is sub-standard. So, "good" broadband service (option for 25 MB or better) is generally a cable monopoly, and a hodge-podge of FiOS (25m households passed), a steadily improving U-Verse offering from AT&T, a handful of regional players such as RCN, and a pixie dust of Google Fiber if you've won the lottery and live in KC or Austin.

In wireless, as we close in on four national LTE networks and an improving spectrum situation, the U.S. is in the top tier of countries from the standpoint of number of competitors, average mobile broadband speed, and monthly data consumption (though not price), Masayashi Son's derogatory comments about the industry notwithstanding.

He's right, though, when it comes to broadband. Here, the U.S is in middle of the pack. There are many countries who are behind the U.S. on most economic indicators but are well ahead of us with regard to broadband competition, speed, and price. Broadband is sort of like our health care system--we spend a lot but don't get that much return for it.

We also spend a lot, comparatively, on connectivity. The average household spends $400 per month for the cable, broadband, phone, and wireless (assuming three smartphones and one feature phone). That's nearly $5,000, non tax-deductible dollars per year, folks.

So, bottom line: broadband is expensive, not that good, and not that competitive. And it's about to become even more concentrated, with a post-TWC Comcast owning some 40 percent of the broadband market.

Enter Wireless?

Now, let's look at this from a demand and supply perspective.

Consumption. The average smartphone user consumes about 1 GB per month. The average broadband household consumes about 30 GB per month. Households with heavy consumption of OTT video such as Netflix quickly approach 100 GB per month of usage (an hour of HD video is 3-5 GB). Cable companies generally have a "soft cap" of 250-300 GB.

Speed. The average data speed in broadband across the U.S. is about 15 Mbps, according to the FCC. Cable companies, especially, have been steadily improving their broadband offerings. Comcast's "default" broadband service is 25 Mbps, with options for 50 and 100 Mbps.

The average speed in LTE is about 10 Mbps download, where available and where service is good. Many users experience download speeds in excess of 20 Mbps, but this is not consistent.

Pricing. Here's where the delta between the economics of broadband and wireless really shows up. In broadband, the average unbundled bill of about $60 per month works out to $2.50 per GB used ($60/25 GB used), but only $0.25 per GB paid for, using a soft cap of 250 GB. Broadband ISP fees on lower-tier, usage-based plans average about $1 per GB. In wireless, the average retail price paid is $8-10 per GB. 

Now, we have to factor in the likely usage, plotted against the technology evolution.

All the studies focus on the incredible growth in wireless data usage, but we should also consider what that curve looks like in broadband, as well. Video is a huge driver. A family of four with two teens, the usual array of connected devices, and Netflix streaming, is trending toward 100 GB of broadband usage per month. That's the sort of consumption that would have a wireless operator CTO quaking in his or her boots.

And, we're likely to see more of those 100 GB households over the next few years. Broadband is becoming an increasingly attractive component of the cable and telco business. Today, most cable companies actively discourage a broadband-only service, making it relatively unattractive from a price standpoint to unbundle the vaunted Triple Play. But as cableco margins become increasingly squeezed on the content side by ever-escalating programming fees, they are likely to stop fighting the Netflixes and Aereos of the world and might actually market a premium, OTT-centric broadband tier, say $70-80 for 50-100 MB speed and a 200-300 GB cap. Want 4K video? That's an extra $10-15 per month, thank you very much.

Now, how about wireless? Well, the average LTE experience, where coverage is good and capacity readily available, is good enough for most of what people need broadband service for today. 10 MB download speed in a stationary environment delivers reasonably good video quality. Using the full bag 'o tricks (carrier aggregation, 20x20, high-order MIMO, small cells), operators can probably deliver 20-30 MB pretty reliably, with a path to 50 MB, especially in a fixed environment, and special CPE in the home.

So the problem in wireless is not with throughput. It's with capacity and economics. I assume it costs an average wireless operator $1 per GB to deliver data. Plot that against the 25 GB average household broadband consumption, and that works out to about 10x what that operator collects today from a typical mobile data user ($1 per GB instead of $10 per GB).

The other issue is capacity. The capacity of a 4G cell site today is in the range of 100 MB-500 MB of simultaneous usage. The full bag 'o tricks roadmap takes us up to, perhaps, 1-2 GB per site in a couple of years. That works out to 150-200 simultaneous streams at a 7 Mbps per stream requirement for HDTV.

The elephant in the room for wireless is video. In order to offer a competitive broadband service, wireless operators are going to have to find a way to deliver video more efficiently and economically. There are numerous possibilities here, from content caching to more intelligently managing where to put the content in the network. But there's no magic bullet. Wireless operators and equipment OEMs will have to develop closer partnerships with companies such as Netflix and Amazon, in order for everything to come together.

Bottom Line

So where do I net out? Given the totality of demand for data--all the devices, mobile and fixed use, plus a healthy amount of video thrown in--I don't believe wireless can compete effectively in broadband, for a typical household. What I think we'll see a more structured segmentation of the broadband market. The "Netflix" households are gonna need their DOCSIS- or fiber- based broadband services, with 50 MB speeds and oodles of capacity.

But if you don't need a world-beating, video-centric broadband service, wireless will start to become a viable alternative. There are a lot of price-sensitive households that have a hard time swinging circa $400 per month for connectivity. They would love a solution where they don't have to pay for separate fixed and mobile broadband. I think there's room for a fixed wireless service to the home, delivering throughput in the 10-20 Mbps range, for $30-40, with perhaps a 25 GB cap. I think the addressable market for this type of service is about 1/3 of U.S. households. This sort of harkens back to how MetroPCS and Leap Wireless looked at the market when they started going after price-sensitive users who didn't want/couldn't afford to pay for both landline and cellular voice. It might play out like this in broadband.

Mark Lowenstein, a leading industry analyst, consultant, and commentator, is Managing Director of Mobile Ecosystem. Click here to subscribe to his free Lens on Wireless monthly newsletter, or follow him on Twitter at @marklowenstein.