Dirty, rotten scoundrels

Felix is far, far too kind to the telcos of the world, when he writes in an otherwise excellent post  (picked up by Krugman), of the lack of decent network coverage in NYC:

If I were an AT&T shareholder, then, I’m not sure how much money I’d want my company to spend on beefing up 3G wireless in New York and San Francisco, especially when there’s little obvious return on such an enormous investment. Sometimes you make more money with cheaper unhappy customers than with more expensive happy customers. And this could well be one of those times.

Firstly, let Baruch share his amusement at being cited as an expert in the physics of radio networks; he is at heart a poncy arts student, and is more likely to think an Erlang is a seminal German film director than the measure of network whatever-it-is that it is.

Secondly, and more importantly, western telco operators are in fact hapless, scrooge-like con-men, who have consistently misled their investors and customers about the extent of their next generation network investment. They have scrimped on 3G spending for years, preferring instead to provide investors improbably fat, 10% plus, free cash flow yields, and 5-8% dividend yields. As of this point, 10 years after getting the spectrum, they have still not properly built out their 3G capacity; and now the chickens, in the form of iPhones, have come home to roost.

There’s a lot of spin about this. Operators say they have spent what they need to build out robust networks. They say that 3G equipment is cheap, and they haven’t had to spend lots of money to get decent capacity. Given that they control all the data, it’s hard to refute this claim, except for the fact we iPhone users know it’s not true. Here are some key facts: China has just started its 3G rollout. In 2009 Chinese operators installed more 3G base stations in one year than western European operators have, ever. Other than trials, the Chinese haven’t had any 3G subscribers to speak of. I used to think that spoke of a certain extravangance on the part of the Chinese; now it looks like the weird ones are the western operators. NTT Docomo’s (the largest wireless carrier in Japan) capex to sales ratios have consistently been 30-50% higher than Vodafone’s (the biggest European wireless operator), which have hovered around 10-12%. And Vodafone has been the least sinful European operator, the one that built out more 3G than any other.

It’s difficult to know exactly how much AT&T has been spending on wireless capex in the past few years; they did a big merger with Bell South a couple years ago, and also have a big fibre project going on where they are rewiring many of their fixed line properties. It’s much easier to compare 100% wireless players like DoCoMo and Vodafone. For what it’s worth, AT&T’s capex-sales ratio was weirdly high 17% in 2008, but was 13.3% last quarter, and closer to that for most of this decade. Verizon has tended to spend much more, consistently about 18% since 2006 as well as last quarter, but again, the waters are muddied by a big fixed line project.

What IS clear, however, is that last Q AT&T gave me, their investor, a whopping 15% FCF margin and FCF yield in 2008, and pays me 6% in dividends every year. Now 6% for AT&T  is pretty decent when the 10 year pays me 3.8%. You, dear iPhone user, are giving me that by paying full whack and suffering a craptastic service. Ask the call centre about this next time you ring up to complain.

Now, in their defence, operators like AT&T and Vodafone haven’t needed to have extensive 3G networks for most of the decade because no-one was using 3G for data, unlike in Japan. Rather 3G was an overlay on networks for voice, relieving presure on 2G networks that were getting full in dense urban areas like London and New York. It was great for this, you only needed a limited amount of 3G for a lot of voice capacity. Remember also that for most of this decade, operators were strapped for cash. They had massively overpaid for spectrum and ill-fated forays into internet businesses during the bubble, and needed to recapitalise.

The other factor leading to underspend was that the period when 3G data usage really took off, the past 2 years, coincided with Our Recent Difficulties. During the credit crunch companies stopped spending on almost everything, telcos included. As their revenues fell, operators had to cut back on capex to keep their FCF and dividend yields up, and be the defensive stocks all the analysts and investors told them they should be. Given the wireless broadband revolution that was underway, this was precisely the opposite of what they should have been doing; in a normal economic environment it would have been a time of increased capex.

This only goes to understand the telco operators, not to forgive them. However you want to look at it, they have failed to future proof their networks, which is of course, all they have to sell. They screwed themselves; they peed their own pants. They fully deserve the PR disasters which have befallen them. And for some, it’s not just going to be a PR issue.

Not all operators have messed up as bad as others. CDMA-EVDO networks like Verizon’s are broadly more efficient than GSM-WCDMA networks, like AT&T’s, and Verizon’s network isn’t chockfull of iPhone facebook app users downloading high bandwidth photos of each other. At some point this year Verizon is going to get a CDMA iPhone; suddenly all the frustrated iPhone users on AT&T will have an alternative, and there are likely a lot of 2-year contracts coming up for renewal this summer. Indeed, I’d guess lots of potential customers have been waiting for a Verizon iPhone precisely because they didn’t want to deal with the AT&T network. It’s likely to be another big jump in iPhone penetration.

So here’s a case where starving your network really IS bad for business. It is true, as Felix writes, that  “sometimes you make more money with cheaper unhappy customers than with more expensive happy customers”; but in a competitive market that window of opportunity doesn’t last very long. AT&T, and probably every other western operator, is probably going to be spending a lot more on wireless capacity in the next few years.


10 thoughts on “Dirty, rotten scoundrels”

  1. Here’s an illuminating quote from Ivan Seidenberg, CEO of Verizon, from 2005:

    “Why in the world would you think your (cell) phone would work in your house?” he said. “The customer has come to expect so much. They want it to work in the elevator; they want it to work in the basement.”

    Up until recently, carriers viewed the primary use case for cell phones as mobile and outdoors. When operators plan their radio network, they have a concept called “area reliability”. Area reliability is the fraction of the area covered by a cellular network where service is actually available. The typical target value is 90 or 95%. For mobility networks. this was deemed acceptable because if someone can’t get service, then they can simply move to somewhere where they can.

    Needless to say, this assumption doesn’t hold true in a world where people are cutting the cord and using their cell phone as their only phone, at home (in basements even!). The carriers have evolved somewhat to recognize this fact but they still retain the view that it’s unreasonable for people to expect ubuquitous coverage. This is largely driven by their conceptual model of the radio network architecture. Cellular base-stations are expensive to build and operate, requiring high DC power, battery backup, air conditioning/heating, etc. The capital and operational costs are high and therefore limiting the number of sites is critical.

    Your point about how carriers have viewed 3G is spot on. It’s primarily been seen as a means to increase voive capacity and as such, it’s been deployed in a way that’s wholy inadequate for the traffic tsunami created by the iPhone and its wannabees. When carriers do their RF planning, they typically have a target minimum data rate that all subscribers in the coverage area can achieve. For 2G voice, this data rate is about 10 kbps (for both GSM and CDMA). For superphones, it’s a few hundred kbps in the uplink. However, the maximum transmit power of superphones isn’t any higher than 2G phones. The higher data rate requirement reduces the cell size dramatically. If you also assume a target area reliability of 99% (instead of 90 or 95%), you’re talking about 10 times as many base-stations. Additionally, you can no longer get away with a 1.5 Mbps T1 as backhaul to a base-station. You need something more like 50-100 Mbps.

    So to summarize, you need approximately 10 times more sites and and approximately 50 times more backhaul, per site, in order to have a network that can handle the new world of superphones with real internet capabilities. If you retain the standard mobility network architecture, this works out to an inconceivably higher CAPEX and OPEX than is feasible. The fundamental problem is that squaring this circle requires serious innovation in the network design and the big North American carriers aren’t exactly hothouses of innovation. They’ve seized upon femtocells because they dangle the possibility of solving this problem with no actual effort on their part. Make the customer, who has “come to expect so much”, pay for the network infratructure that’s really the carrier’s responsibility to provide.

    1. Ivan Seidenberg, CEO of Verizon, from 2005: “Why in the world would you think your (cell) phone would work in your house?” he said. “The customer has come to expect so much. They want it to work in the elevator; they want it to work in the basement.”

      Ramster: “Up until recently, carriers viewed the primary use case for cell phones as mobile and outdoors.”

      That is shocking. How could a CEO be so ill-informed about his own industry? Does he even use a cell phone? It must be nice to “compete” in an industry where you have the luxury to treat consumer demand as an annoyance instead of desperately responding to it lest you become extinct.

  2. @Ramster? Why has the femtocell push been so lacklustre though? Is it a technology issue? Or are the network cores not up to it?

  3. Ramster, fact is the Japanese manage it. They just spend more in capex. And backhaul’s pretty cheap and is a mostly one off cost.

    They can do it, AT&T and VZ, if they really try, and in fact, I suspect they probably will; they’ve been spending billions on fiber to the X, rewiring their fixed line networks. That’s coming to an end soon, and they can probably maintain current FCF yield and dividends while spending on one big project. They can’t do big spends on both fixed and mobile AND pay us loads of money.

    Meanwhile, while they are waiting for the fixed line projects to end, they have to work with crappy wireless capacity and AT&T in particular will have to put up with the brickbats.

    Anyway, that’s what I was thinking. Am probably wrong, but let’s see. Thanks for the comment, it was very interesting.

  4. I’m an European fan of this blog and it is a good but plain obvious post.
    Visiting San Francisco conferences every year is always a horrifying experience how bad wireless networks are in the US, and the capital of Silicon Valley is no exception.
    I habitually joke about having no coverage around Market street and need switching operators just to be able to download email headers on my smartphone.
    I haven’t seen this kind problems ever in Europe although I live and work in big metropolitan areas (much bigger than SF).
    The most hilarious joke is – which my European friends simply can’t beleive – many US users still pay for incoming local calls. (This was stopped here around 1994 when GSM networks were introduced.)
    And have I told you my local operator (T-Mobile) wanted to install my own femtocell free of charge to make sure I’m happy? (I said no, thank you, the coverage is fine )

  5. It’s easy to see why the operators love the femtocell concept. It allows their view of how wireless networks should work to revert to the outdoor only model (though it still doesn’t address the fact that outdoor only coverage with 3G and 4G networks still requires a lot more base-stations due to the lower cell radius required by the higher uplink data rates of smartphones). The indoor coverage problem is solved in an ideal way, one that costs the carriers zilch. The customer pays for the infrastructure (i.e. the femtocell) and the backhaul (their broadband internet connection). All the carrier provides is the wireless spectrum, which they already have!

    However, there are a few flies in the ointment. The first is a technical problem and the second is with the business case.

    The technical problem is related to radio signals and their tendency to go where you don’t want them to (and also to not be where you want them). A given carrier’s cellular networks uses a fixed amount of radio spectrum, which is re-used in every cell (or group of cells). The signal used to convey user’s traffic in one cell looks like interference to neighbouring cells. Minimizing this interference is a key aspect of how the wireless network is planned and a key mechanism is to limit the power of your transmitters. However, you’re also trying to punch your signal through walls and past obstacles like hills, foliage, buildings, etc, which is best achieved by cranking up the power. Balancing these contradictory requirements makes radio network planning a complex exercise. In a given city, a 3G of 4G cellular network may have a few hundred base-stations. Widely deployed femtocells may results in hundreds or thousands of times more base-stations, all using the same frequencies. Though the femtocells are individually much less powerful than an outdoor base-station, their cumulative power can increase the aggregate interference level, which in turn makes the performance of the outdoor network worse. This is a very tricky problem and there are technical solutions but the absence of widespread deployments makes it a risky proposition. This is one of the reasons that the femtocell solution has been rolling out very slowly. The conceptual framework of cellular radio networks has been around for 20 years and is reasonably well understood. Femtocells are a radical departure from this framework and hence carry significant risk.

    The business case problem has a few dimensions. The first is simply what’s in it for the consumer (or enterprise)? Why would I pay for a device that does something that the carrier should be doing already, provide coverage in my premises. If they’re not doing that, what am I paying for? Perhaps if the femtocell was free and I get a discount for paying for the backhaul (i.e. my broadband connection). The other problem is related to the backhaul. What happens when my broadband service provider is different from my wireless carrier? They may look at the femtocell application as outside the terms of their service agreement. Or they may deliberately throttle the service in some manner unless they get their cut. Also, the cellular core network (which connects all the carrier’s base-stations to the public internet and phone network) is typically very tightly designed with rigorous performance requirements, in particular w.r.t. latency. A generic broadband connection may not be able to provide such guarantees (particularly when the network is congested), which in turn could degrade voice quality.

    Ultimately, femtocells have to be compared to the alternative, wifi. What does a femtocell do for you that your home wifi doesn’t? For internet applications, nothing. Any decent smartphone these days will use wifi if its available. The carriers originally hated wifi capability in smartphones because it allowed the user to bypass their network, and hence control. The first Blackberry Storm didn’t have wifi capability because Verizon demanded that RIM omit it (there was no technical reason). It’s a bit trickier with voice but there are solutions out there. and they’ll get better (think hybrid VoIP over WiFi and cellular calls with the user not knowing or caring which one is in use and seemless switching when necessary) The reality is that innovation in the handset and application space will outpace innovation in the carrier’s network and business models. All the carriers can do in response is try their best to impede this process (i.e. stop Google Voice).

  6. Baruch, regarding the Japanese (or Koreans or many Europeans), that have a key enabler…widespread high speed wired broadband. This provides a pre-existing infrastructure for the immense backhaul requires for 3G and 4G wireless networks. So to a significant degree, US wireless carriers are constrained by the crappy state of wired broadband in the US. Clearwire is being forced to use a ton of wireless backhaul (point to point microwave links) for their WiMAX network precisely for this reason.

    1. Oh aren’t they poor little woogums. Fact is Ramster, AT&T and VZ, the big wireless guys in question, are also the wireline guys; so the crappy state of wired broadband in the US is precisely the fault of the umm, wireless guys as well. And their brain-dead cheapskate decision to use those thingummywachits from Tellabs for wireless backhaul instead of some decent, future-proof fibre tech.

      T-Mob doesn’t seem to have the same problem. Or maybe they just haven’t got anyone using their network.

      No, all it needs is determination, and above all money, which washes all sins away. I reckon they’re going to spend great gobs of it.

  7. Baruch, no arguments there and I certainly didn’t mean to imply that I’m defending these dinosaurs. As wireline and wireless operators, AT&T and VZW have the least possible excuse for their crappy backhaul. I’m not sure that the other operators are really that much better than AT&T; they just haven’t been hit by the iPhone generated traffic onslaught. All the new superphones out there will change that.

    1. No, I know you didn’t. I just wanted to rant.

      Actually, looking at all these super phones out there at CES, have you checked out admob monthly reports of browsing by OS and by phone model (they’re free)? It’s terrible. Everyone other than iPhone has a highly asymmetric struggle on their hands.

Comments are closed.