Flattening the Smartphone Market

Common cellphone operating systems like Android, WebOS, and Maemo are depriving cellphone carriers of one of their most treasured means of keeping customers in the dark and feeding them bullshit. They’re making smartphones comparable to each other, and by doing so brutally intensifying the competitive pressure on the carriers.

Before these common platforms, one of the ways carriers had to quell customer demand for features like tethering was simply that the feature sets of different phones were difficult for customers to compare in meaningful ways, and there was no real benchmark for what a cellphone could or should do (I mean, other than make phone calls).

That’s all changed now. To see how, consider the impact of the Froyo announcement – Android 2.2. This was a feature list of what upcoming cellphones would be able to do that wasn’t censored by a carrier. Every actual Android phone offering from now on will be compared against the Froyo feature list and against all other Android cellphones. And if any come out with an Android feature disabled or available only at extra cost, the pressure won’t be on Google – it will be on the carrier.

Consider as an example the feature of 2.2 that looks most deadly to the iPhone right now – its ability to act as a WiFi hotspot for up to 8 devices. If you’re Sprint, and you’re betting your company on fielding a 4G data network that means always-on high-bandwidth Internet for your road-warrior customers, this is great news; you’re ready, and the marketing for the EVO 4G is already promising it. If you’re not Sprint, your data network is probably underprovisioned and this feature scares the living crap out of you.

But what are you going to do about that? If you disable the feature, you drive your customers to Sprint. If you charge extra for the feature, you still drive your customers to Sprint (people love flat-rate plans; they have a psychological resistance to being nickled and dimed, even when it’s an economically rational choice for them). In effect, Sprint is defecting from the feature-suppression cartel that has kept tethering out of the hands of almost all U.S. cellphone users. And this is difficult for the other carriers to obfuscate because now, everybody knows that Android 2.2 is supposed to do that.

I’ve written before that the carriers are losing control of their cellphones’ feature sets to Google. First to go has been the ringtone market; not much life left in it when you can grab any random MP3 or OGG off the net and use it as a ringtone. Android 2.2 is where this becomes overt. There’s been a lot of talk about how Apple will have to play catchup with the 2.2 feature list at its iPhone 4G announcement, but the story behind that is that the carriers will now have to play catchup. The same competitive logic that applies to tethering in 2.2 will apply to any other feature Google chooses to ship in future releases – and consider the havoc this could wreak if it were (say) VOIP support or SIP conferencing.

Back to Apple. Anybody want to bet that Steve Jobs hasn’t already been on the horn to AT&T demanding that they allow the iPhone 4G to ship with tethering enabled? From his point of view this is now a must-have if the 4G isn’t going to look like a weak second-best. But AT&T is going to push back, oh yes they will, because they’re already notorious for iPhone service problems stemming from network underprovisioning and congestion. Hilarity, and possibly some bloodletting, will ensue. There are only three possible outcomes here: (a) iPhone 4G fails to ship with tethering, in which case Apple and AT&T both take a hit, (b) iPhone 4G ships with tethering and it sucks, in which case AT&T takes a hit and Apple may dodge a bullet or not, or (c) AT&T ponies up a gigabuck or three to upgrade to 4G at record speed.

The punchline is that Google is now setting the agenda for Apple, AT&T and the carriers in general. Products are now comparable; everyone has to react to Android feature announcements. Google can probably raise the bar faster than they can react. What if it were to, say, announce native support for SIP teleconferencing in 2.3? Suddenly proprietary videophone support wouldn’t be a differentiator anymore.

It’s by steps like this that Google will hammer the carriers to their knees, consigning them to the role of low-margin bit-haulers while Google ratchets up the smartphone feature set. I’d say it has already begun, but that would be understating the case; with Android 2.2 the process is already quite advanced.

86 comments

  1. I have to say this is right.

    “But what are you going to do about that? If you disable the feature, you drive your customers to Sprint. If you charge extra for the feature, you still drive your customers to Sprint (people love flat-rate plans; they have a psychological resistance to being nickled and dimed, even when it’s an economically rational choice for them).”

    Sadly, that is what Sprint is doing with tethering though. They are charging an extra $30/mo for the feature. Hopefully that will go away when someone offers some competition.

    ESR says: Right, and the first carrier to ax that surcharge will pressure the others into doing it too.

  2. Uh huh. You seem to forget that the iPhone has had tethering since 3.0. It’s available internationally, and apparently works great. But it’s presence hasn’t pushed AT&T to budge. Instead, people blame Apple for not having tethering.

    Ironically,. AT&T allows tethering on other devices, but not the iPhone, because the iPhone is so easy to use that they’re afraid that people other than non-techies may actually use it. We’ll see at WWDC if AT&T starts allowing tethering, I suppose. Alternatively, perhaps the iPhone is coming to Verizon and Sprint as rumors indicate, in which case we’ll have tethering on one or both of those networks, pressuring AT&T to allow the same.

  3. (d) Apple ponies up a gigabuck or three to buy out whatever’s left of their AT&T contract and ships with Verizon (or for a real “one more thing” shocka, Sprint…).

    1. I agree Apple has to try to go multicarrier as soon as it can get out of its exclusive with AT&T. But I wonder if Jobs is going to find any willing partners.

      Even assuming Apple could get legally free to do it by WWDC, I think it’s not gonna be Sprint, not as hard as they’re pushing the EVO 4G. It would be nuts to confuse their customer base, complicate their tech support, and multiply their SKUs that way.

      In the longer term, look at it from the point of view of a carrier exec with a brain in his head. If you’ve got a choice between iPhone and Android, why are you going to pick iPhone? As bad as losing control of your feature set to Google looks, taking diktats from Steve Jobs has got to look worse; don’t-be-evil sure beats I’m-as-evil-as-I-wanna-be. Google won’t try to skim the cream off your handset profits, because it doesn’t need to – it’s in the advertising business. The days when Apple could claim its “superior experience” infallibly drew customers away from Android ended with 2009. So, er, what’s the iPhone value proposition again?

  4. esr wrote: “And if any come out with an Android feature disabled…

    Is Android not open source? How is it possible to “disable” a feature if one can just recompile it and load it back on the phone?

  5. Sadly, that is what Sprint is doing with tethering though. They are charging an extra $30/mo for the feature. Hopefully that will go away when someone offers some competition.

    With a few minimally technical steps you can tether your Android device using AziLink and openvpn, and you don’t even have to root your phone (but you still do need a USB cable). As a middle finger to Sprint’s tethering policy it works acceptably well.

    As bad as losing control of your feature set to Google looks, taking diktats from Steve Jobs has got to look worse; don’t-be-evil sure beats I’m-as-evil-as-I-wanna-be.

    Granted, Teh Steve is looking more and more like Hank Scorpio these days, but Cypress Creek is such a nice community, and Globex is such a fun place to work! When it comes to the experience your user base sees, you could do a hell of a lot worse than blindly following Steve Jobs. Remember when iPhones didn’t even have apps: just the browser, email, YouTube, and whatever things were preloaded by Apple, and they were still awesome? This was when Windows Mobile still owned the smartphone market, and delivered a much broader feature set.

  6. Is Android not open source? How is it possible to “disable” a feature if one can just recompile it and load it back on the phone?

    You would have to root the phone first before doing that, which not a lot of people are capable of doing or inclined to do.

    The Backflip, an entry-level Android device available on AT&T’s plan, ships with installing non-Market apps over the Web disabled. (I think you can still enable USB debugging and install apk’s over the USB cable, though.)

  7. I still can’t see Apple making a CDMA iPhone. Talk about SKU proliferation. Right now, they get the benefit of making one SKU that will work worldwide, and having only one hardware platform to work with.

    If they get out of their AT&T exclusivity deal before its 2012 expiration, it’ll be to go with T-Mobile as well as AT&T.

    I still don’t think Android’s open-sourcedness is all it’s cracked up to be. What does someone with a Droid have to do to rebuild and reload? Is it even possible?

  8. “So, er, what’s the iPhone value proposition again?”

    Customers seem to prefer it. Personally, every person I know who owns an android doers so because they wanted an iPhone but hates AT&T, my wife included. In fact the only market where the two currently coexist is AT&T, where the iPhone outsells android offerings by some ungodly ratio. We’ll see if the same thing happens when or if the iPhone if ever offered on Verizon.

    1. >Customers seem to prefer [the iPhone].

      After last quarter’s unit-sales numbers this claim is at best highly dubious, but let’s assume for the moment that it’s true.

      It still won’t necessarily be determinitive from the point of view of a carrier strategist. The customer preference for iPhone would have to be strong enough to translate into profits large enough to offset the fact that Apple takes the top off your handset revenues and utterly controls your phone OS. Good luck with that…

  9. RE: Tethering on Sprint

    Prior to the googly phone I’m running now (the Moment) I used Palm Treos/Centros and all of them were capable of tethering with the software that Sprint and AT&T shipped. It wasn’t terribly fast (because the phones weren’t 3G) but it worked and I used it when I had no choice.

    I’m wanting the wireless tether on my Moment, tried using Barnacle, but Alas! it doesn’t like the Moment with Android 2.1. Scattered reports of “success” appear to be overrated.

    -sigh-

  10. > After last quarter’s unit-sales numbers this claim is at best highly dubious,

    Problem is you haven’t controlled for the network. Many of those buyers wanted Verizon first and a smartphone second. Were the iPhone on the Verizon network, undoubtedly some number of those people would opt for the iPhone. We just don’t know what that number is as of yet.

    > It still won’t necessarily be determinitive from the point of view of a carrier strategist.

    So what you seem to be saying then is that the iPhone won’t migrate to other carriers because android offers them such a better deal. I could believe that. I’m arguing that if it does move then it will likely be very competitive, and you may well see the same kinds of results that you see within the AT&T ecosystem, namely an iPhone domination.

  11. Well, it’s just possible that, given the chance, carriers would drop Android and go with the iPhone. There are a lot of telecom executives who absolutely do not want to be commodity bit haulers, and who very much want to increase their revenue per customer with exclusive features. The market may correct for that, but given the cost of infrastructure buildout as both barrier to disruptive competition and an incentive to to increase their revenue per customer, maybe not.

    Yours,
    Tom

    1. >There are a lot of telecom executives who absolutely do not want to be commodity bit haulers, and who very much want to increase their revenue per customer with exclusive features.

      And, at this point, they’re fucked in either direction they jump. They can’t do that anymore in any but the most trivial ways. Either Apple is going to control their feature set in the name of “user experience”, or Android will do it with pre-emptive announcements like the Froyo thing. The carriers can desire things to be different, but they’re not going to get their wish, because the only way for them to do that would be to roll their own smartphone OSes that are better than Android or the iPhone OS. Nobody believes that can happen at all, let alone with acceptable time to market.

  12. Problem is you haven’t controlled for the network. Many of those buyers wanted Verizon first and a smartphone second. Were the iPhone on the Verizon network, undoubtedly some number of those people would opt for the iPhone. We just don’t know what that number is as of yet.

    I often hear iPhone fans try to justify lower-than-expected iPhone sales with claims like “well it’s because it’s only on AT&T and not [Verizon | Deutsche Telekom, err…T-Mobile | etc. ]….” Y’all are aware that though Verizon and AT&T are currently #1 and #2, historically iPhone has been very successful in converting Verizon customers to AT&T customers, right? Also, the difference in market share has never been very great — Verizon current commands a 31% share to AT&T’s 25% share. And, historically, they’ve been neck-and-neck like that.

    What I’m getting at here is that carrier choice is a very fickle thing. Wireless subscribers are often willing to jump ship to a different network to get a device they want, to get a better price or plan, or any number of carrots that the wireless carriers use to entice each other’s customers. It’s called churn, and the carriers hate it. That iPhone is tied to AT&T hasn’t been a death sentence for it — up until now. From a non-fanboy, non-geek, Joe Sixpack point of view, the differences between Android and iPhone aren’t enough to justify picking one device over the other. But if Joe Sixpack can get either a better price point, better phone plan, or better call quality or coverage from a different wireless provider, then, yes — that’s when iPhone’s provider lock-in becomes its death sentence.

    I’m arguing that if [iPhone] does move [off of AT&T] then it will likely be very competitive, and you may well see the same kinds of results that you see within the AT&T ecosystem, namely an iPhone domination.

    It may be a moot point, but I doubt it. When people start comparing phones and capabilities, most are going to go with the best bang for the buck And in the new phone ecosystem Google is creating, that means that iPhone will stay a minority player, just like Apple’s products in the PC business. History often repeats itself, and I remember arguing this same stuff with Apple kool-aid drinkers in 1984-1986, who were all certain that the Macintosh would dominate personal computing.

    1. > History often repeats itself, and I remember arguing this same stuff with Apple kool-aid drinkers in 1984-1986, who were all certain that the Macintosh would dominate personal computing.

      That’s right. The belief that “better user experience” is a recipe for market dominance simply isn’t borne out by history.

  13. Might I suggest:
    “http://www.mobilecrunch.com/2009/06/19/android-users-are-stickier-than-iphone-users/”

    This seems to be the best “hard(ish)” data I could find on where Android and iPhone stand in the market. Couldn’t find anything that specifically deals with the number of converts from one camp to the other (anybody seen anything like that)?

    Of course, I’m wondering if the FanBoys will be telling how “Apple Rules” when the 4G phone comes out and kicks everybody’s butt in sales because of the upgrades for the next 2 quarters.

    My take: It’ll be iPhone’s last bounce before settling down to a slowly bleeding users to Android, in the same way that RIM has been bleeding for the last 12-18 months.

  14. That’s right. The belief that “better user experience” is a recipe for market dominance simply isn’t borne out by history.

    Market dominance was never Apple’s plan. Being the guy whose taillights the mainstream market players chase is closer to the mark. It means you’re where all the exciting, innovative stuff happens.

    Joe Sixpack is constrained by price/quality economic tradeoffs; Apple’s goal is and (under Jobs) always was to deliver the best platform. This they achieved with iPhone.

    A sure-fire recipe for market dominance is to be 100% backward compatible with whatever installed base is out there while offering an upward migration path. This is why Windows and the crack-headed x86 ISA won. Not that backwards compatibility isn’t a concern for Apple; it’s just a secondary one. Steve will move mountains to deliver a better platform, and then he’ll erect an enormous sound stage backdrop with mountains painted on in their exact prior scale and location, just to placate those who relied on them being there.

  15. My take: It’ll be iPhone’s last bounce before settling down to a slowly bleeding users to Android, in the same way that RIM has been bleeding for the last 12-18 months.

    The only users Apple is in danger of losing in appreciable numbers are the geeks who are butthurt over their recent developer policy changes.

    They will have a proportionally smaller piece of the overall smartphone pie, as more affordable low-end/midrange Android devices (like the Backflip) become available and the market expands to an increasing number of Joe Sixpacks who must make quality/price tradeoffs.

  16. Umm, haven’t phones always had common operating systems, so what changed now? Is it that Google is forcefully devoting more resources to Android and accelerating change? If so, why was Verizon still able to try and cripple the Nexus One and then abandon it? I’m skeptical of any analysis that invokes competitive dynamics for the cellular market as that market is an oligopoly with at most 4 carriers in most markets. If anything, the rush of consumers now entering the previously mostly enterprise smartphone market is probably what is accelerating investment and change somewhat, but the carriers are adept at using their oligopoly status to stonewall change. As for your silly assertion that consumers have some sort of in-built psychological aversion to metered plans, if that were true, billing of water, electricity, gas, and phone utilities would all be dominated by flat-rate plans by now, but magically they don’t. The only utility that does is landline internet and they are trying to move away from that model also. The real issue for consumers is having real-time accounting that allows for predictable bills, which are now trivial to implement as bandwidth monitoring apps on all phones.

  17. Jeff Read: Problem for Apple is, the phone market ISN”T the PC market with it’s long development lead times. As Android pushes more features, the UI improves, the hardware becomes cheaper, and they become more carrier agnostic (which they already are) with respect to features, MOST people will choose to avoid the Apple lock in. And the GUI isn’t sufficiently different between the two that people will hesitate to make the jump (hell, most people are USED to learning a new phone every 12-24 months).

    So, Apple can only retain it’s lead technologically by being the coolest. But then that will become a problem too as the development cycle of Android apps continues to ratchet up (it’s already in the weeks for most apps, and in the months for the OS cycle) and THAT’S where Apple looses the coolness race. For every cool feature Apple puts out, Android will put 3 or 4 on the pile. For every performance improvement iteration that Apple goes through, Google will do 2 or 3. Apple cannot keep up, it’s simply not possible.

    I’m not predicting the demise of the iPhone, but I am predicting the demise of Apples’ margins on the product as they slip into the same place that has become so familiar for Palm over the last decade. In the end, there will be people who buy the iPhone because of the logo, same reason there are lots of people who buy Cisco or Microsoft or IBM even though there better alternatives for MOST jobs. Brand image counts for a lot with consumers, but your argument that Apple will lead the market seems… off a bit. They are technologically neck-and-neck with Android now, they had a 2 year head start, and a HUGE installed base to leverage, and they’re panting already.

    I have a feeling that Google is just unwinding, and as I pointed out earlier (in a previous entry) and ESR comments on above, Sprint is not a victim here, they are a willing accomplice to Google’s plans. They’re gonna ride Android the way AT&T rode iPhone. So far Google’s just been getting “feature complete” with iPhone. Now, if what the 2.2 announcement says is true, Google’s moving into the drivers seat. It’s Google’s race to lose, and I wouldn’t bet against them.

    As for being “butthurt”, with the growing market share and broad carrier coverage, professional apps developers would be crazy to ignore Android as a target platform. If some devs are vocal about disliking Apples terms, well that’s their prerogative. It seems silly to me to abandon the iPhone platform, but then it’s likely the complainers are either in the class that is getting booted or squeezed, or they’re not making any money/have a limited audience on the iPhone anyway so it’s not a hard decision for them. “Terms” are usually a convenient excuse to do what you were going to do anyway (kinda like all the companies that went out of business after 9/11, when they were on the verge of bankruptcy anyway :^).

  18. Ajay> As for your silly assertion that consumers have some sort of in-built psychological aversion to metered plans…

    Uh, it doesn’t take but one or two $300 cell phone bills to get the idea that $99 all you can eat plans are a beautiful thing. Or better yet, $55 (Cricket), or better yet $45 (Pocket).

    There was once a time I would have agreed with you, but then I left my local calling area for a day and had a family emergency. I switch the day after I got the bill and never looked back. I KNOW what my bill’s gonna be, I don’t have to check the meter.

    Of course it helps that my employer now pays the bill, but he too uses the $99 all-you-can-eat plan for myself and anybody else with a smart phone. The predictability is worth it.

  19. dgreer, all you’re really doing with such an all-you-can-eat plan is trading billing unpredictability for the dropped calls and other problems on the resulting overloaded network. Perhaps that’s not as big a deal for cellular voice, given how little bandwidth audio uses and the carriers’ currently ridiculous pricing (particularly for texting), but it is a big deal for data. Now, there will always be flat-rate quota plans, where you pay $40/month for a quota of 20 GBs of cellular data, but metered pricing is going to be the dominant model for most online services and unlimited plans are a delusion. I don’t want to use the cellular network that is swamped by teenagers video conferencing each other from their tablets and I’m pretty sure you won’t want to either. ;)

    1. >unlimited plans are a delusion

      Sadly for you, many of us have been living in exactly this “delusion” for years. The fundamental thing to get about the economics of network provision – wired or wireless – is that a network’s costs scale in a way coupled only weakly with bandwidth but much more strongly with the number of techs you need to run it. That number rises on a different (lower) curve, and providers who try to charge as if costs scaled mostly with bandwidth get their customer bases eaten by competitors who are willing to hug the lower curve more closely.

      I should blog about this sometime. It’s a common fallacy that needs to get busted.

  20. esr, I see, how long have you had an unlimited data plan for, particularly on a cellular network? ;) As I noted, most utilities have always used metered billing. The fallacy of those asserting fixed or weakly scaling costs for data networks, who you appear to have joined, is that those are the only factors for price. Price signals are much more about rationing resources, whether it’s electricity or peak usage on a cellular network, and financing further network upgrades with the profits from metered plans. That means that unlimited plans are essentially a scam, as they try to ride some supposed cost curves, while sacrificing QoS or secretly inserting quotas. There is the possibility of innovative solutions, like a Sprint rep recently told me that their 5 GB data limit wasn’t a hard quota: they supposedly prioritize the first 5 GBs of data each month and then any further usage drops to a different service level.

    But unlimited will always be a fantasy because there will always be people who overuse it, just as I currently use up large amounts of bit torrent bandwidth on my wireline network while the grandma down the street uses almost nothing, but we pay the same “flat-rate” price. That works great for me right now, but I certainly don’t want to be exposed to that on a cellular network, where I’m not going to be the one doing the hogging. ;) Ultimately this is a market issue: there are a bunch of people who are economically ignorant who think that unlimited plans are great, without realizing the associated overconsumption that will bite them later. Then, there are people like me who want metered pricing and quotas so that the bandwidth hogs don’t bite us. We can each have a cellular network provide the kinds of plans we want, but because of the superior economics behind metered pricing, it will win out in the end. :)

    1. >because of the superior economics behind metered pricing, it will win out in the end. :)

      Well, if that turns out to be true, it will be a historic first. The direction in which comm-network prices have always tended before is exactly the opposite, because building new capacity is just too easy. Your notion that there are necessarily rationed network resources other than human troubleshooting time is a lot like turn-of-the-century urban planners projecting that their cities would be inundated in horse manure – a superficial linear extrapolation from a very narrow timeslice, and pretty laughable.

  21. The fundamental thing to get about the economics of network provision – wired or wireless – is that a network’s costs scale in a way coupled only weakly with bandwidth but much more strongly with the number of techs you need to run it.

    No kidding. Just ask anyone who’s worked with corporate networks. Your number 1 cost isn’t the equipment, the cable plant, or even the leased lines, but the people who keep it running. Why do people think running an ISP or a wireless carrier is that much different? A network is a network.

  22. Which historic comm-network prices are we talking about exactly, telegraph prices a 100 years ago? ;) Adding capacity to communications networks has never been easy and the problem for networks with flat-rate pricing is that it incentivizes standing still. If you run a metered network, you want to provision capacity as much as possible, because every bit that’s dropped due to inferior QoS is money that you lost. Not only that, you can do the equivalent of night and weekend minutes and only meter daytime usage, pushing more usage to the night (iTunes would just download your movies overnight) and allowing you to do more with less peak capacity. With flat-rate pricing, you have no incentive to build out because your dumb customers won’t pay for it: let them suffer the poor QoS.

    Of course, what hangs over all this is that the telcos are so highly regulated that it actually may work out better for them to choose the inferior flat-rate model, so that they can then go hat in hand to the politicians and demand more taxpayer money. The politicians love grandstanding with dumb economic ideas, so they’d love to say they killed metered pricing and kept flat-rate plans, just like Chuck Schumer’s recent drivel about regulating airlines’ baggage fees. It is fascinating that you support the more socialist solution here, where the buck is constantly passed around by flat-rate plans so as to avoid real economic calculation. It is actually your superficial extrapolation that is laughable as most bandwidth was metered in the ’90s. The ISPs then figured out that as network capacity grew, they could sign up more customers with flat-rate plans, because email and web pages just weren’t a strain on broadband connections and video hadn’t taken off. However, with the arrival of p2p and online video this decade, they’re feeling the consequences of that decision. As soon as the monetization model for online video is figured out and tons of HD video starts flowing, they will either switch back to metered pricing or go out of business, with the metered-pricing ISPs taking over their networks. :) So it is actually you extrapolating wrongly from a tiny slice of history, from the end of dialup metering to the coming rise of video, not me. :D

  23. I am surprised that no one has pointed out the obvious: the Pre ships with the Mobile Hotspot app. Verizon just made it free — it used to cost $45/month. So, what ESR says is already in operation: tethering is a must, and we don’t want to pay extra for it.

    Palm Pre 2: great smartphone, has a world-wide community exploiting the Linux underpinnings of WebOS for all kinds of homebrew apps and patches, and just getting nailed in the marketplace by the marketing geniuses at Google/Android and Apple/iPhone.

    Maybe HP will be able to do more with it. Hope so. I love mine. Really don’t want an iPhone, and not just because of AT&T (although that is an issue.) I don’t especially want an Android either. I like the Pre 2 more and more, the longer I have it.

    Then again, I prefer Fresca to Coke, but can hardly ever find it. Some sort of parallel there.

  24. I think part of the problem is that there’s two dimensions that transfer capabilities can be metered, and people would be fine with one or the other, but not both, because it feels too much like nickel and diming.

    You can basically charge for the pipe, or you can charge for the units through the pipe. Looking at the second case, most people in North America are used to paying per unit of electricity consumed, or unit of natural gas consumed, both of which are supplied by feeds. In almost all cases, the cost to provide a just-big-enough connection and a I-don’t-know-how-to-max-out-this-connection connection. For example, most houses with anything resembling modern electrical infrastructure are rated for 240V service at 150A. Since then it’s gone to 200A and I’ve heard of 300A in some new subdivisions. I’d like to point out that I need to work *really* hard to try and find a way to use that much power for any sustained amount of time. Do you even know the capacity of your gas line (if you have one). In both cases, it doesn’t matter too much because we’ve sized the connection for maximum likely usage, plus a bunch. As a result, we meter (primarily) by usage, not feed capacity.

    Consider, however, traditional Internet usage. I don’t think there is anybody out there which sells a residential package which will meet my *peak* demand. FIoS with 50MB connection comes close, but not quite. Waiting for Linux Distributions or streaming content from Hulu chews up a lot of my pipe. Moreover, we don’t know what the steady-state usage for Internet capacity will be. I think we’re getting close, but we aren’t there yet. As a result, we charge for the size of the pipe. It’s not fast enough, but, here, do what you can with it.

    Metering by both tends to irk people. They get insufficient capacity *and* they have to pay per unit, too. Ick. If vendors/carriers can provide the peak speed people want, then maybe a charge-for-transfer model will work. However, radio spectrum is going to be even harder to provide that level of service on than hard-wired connections.

    I agree that the flat-fee model is going to probably be where things end up once the dust settles, both from a risk-reduction perspective and a customer feel-good perspective.

  25. esr: “It’s by steps like this that Google will hammer the carriers to their knees, consigning them to the role of low-margin bit-haulers while Google ratchets up the smartphone feature set. I’d say it has already begun, but that would be understating the case; with Android 2.2 the process is already quite advanced.”

    If all the carriers are ‘low-margin bit-haulers’ it doesn’t seem like they will have the capital or motivation to build the infrastructure we would like them to have. Ultimately, the customers will have to pay for the network system or it won’t be there. Isn’t this one of those, “fast, good or cheap – pick two” situations?

    And it certainly does not bode well for your friend the marginally employed engineer. Low-margin bit-haulers will likely be staffed by low wage entry level folks. The fast food industry comes to telecom. I hope not.

    1. >If all the carriers are ‘low-margin bit-haulers’ it doesn’t seem like they will have the capital or motivation to build the infrastructure we would like them to have.

      Nah, there’ll still be plenty of money in the business, it’s just that the margins will drop from where they are now. But low margin times a huge run rate still equals plenty of profit.

  26. Which historic comm-network prices are we talking about exactly, telegraph prices a 100 years ago? ;) Adding capacity to communications networks has never been easy and the problem for networks with flat-rate pricing is that it incentivizes standing still.

    Lessee…starting with breakup of AT&T in the 1980s? We went from metered pricing for each phone call to one flat price for all local calls, while the telcos redefined what exactly is a ‘local’ call by creating ‘zone’ calls. Then as cell phones began proliferating and people started replacing their home phones with cell phones in the 1990s, due to wireless carriers giving you flat-rate pricing for all calls in North America, telcos began one creating flat-rate plans that included local, ‘zone’ and local long-distance calls. Long distance phone companies began doing flat rate pricing for their services as well. Finally, beginning in the 2000s as VOIP started taking hold, all the telcos are offering “digital home phone” (read: VOIP) service, also offering flat-rate pricing for all calls in North America. Vonage and other VOIP-only providers upped the ante by adding some international calling in those flat-rate plans as well. Now cell phone providers are offering flat-rate pricing for voice, data, texting, etc.

    So, within recent history, prices have gone from metered to flat-rate in the voice markets.

    The same has been true for Internet Service Providers. Or maybe you just aren’t old enough to remember metered pricing for Internet service?

  27. Ajay: I hope you don’t take this the wrong way, but you’re being unbelievably ignorant here (and I mean that in the nicest possible terms :^).

    After running an Internet Service for the last 15 years, and now running an Internet Service and Data Center belonging to somebody else (God! it’s great not having to handle the billing, collections, etc. ;^) I can tell you that ESR is EXACTLY correct. I’ve seen the model he describes gobble up dialup, web hosting, T1’s, DSL, and it’s now gobbling up fixed wireless (“Wifi”) and mobile wireless (the subject of this article).

    I’ve predicted this for all the other industries I’ve been in, and I cannot see mobile wireless being any different. In the end, it’s just bits, and it’s not significantly more difficult moving 1 bit than moving a million of them (or even a few billion of them).

    The primary difference is in the CAPEX, and when you break that out over time, it scales well. Electricity, Internet transport, etc. doesn’t really matter, because when you get to a certain point (around a 1 Gbps of bandwidth) it’s still possible to get the tier-two carriers to peer with you (assuming you have traffic patterns that pass into their network rather than through their network), then when you break about 10Gbps the tier-ones will find your network sufficiently sized to make it advantageous to peer, or something close to it price wise.

    This is how the Internet works from the business perspective, the costs aren’t linear, they are inversely and exponentially associated with your bandwidth usage. The more you use, the cheaper it gets. The linearity is a fallacy that has been maintained by the old telecom model because it lets the telecom providers continue screwing you linearly. :^).

    ESR says: I understood all that except “CAPEX”. Unpack, please?

    1. >Ajay: I hope you don’t take this the wrong way, but you’re being unbelievably ignorant here (and I mean that in the nicest possible terms :^).

      Yes, he is. I could add “shoddy and biased”, but that would be just mean. :-)

  28. Ajay: I forgot to respond to the QoS issue. That’s simple. If you’re running a major network, and you have transit times > 40ms coast-to-coast, your network is broken. I don’t really care WHAT is on it. Tier-one/two carriers don’t really care WHAT you put on their network, bits are bits. If you want them to classify your bits differently, their gonna charge you out the wazzoo, and you’re still gonna get 40 ms coast-to-coast.

    The only place where QoS is a problem is on a hard-limited, shared network medium (think Ethernet hub here, or wifi AP). This is the case with most wireless implementations, but the congestion is LOCAL, not network-wide, and there’s not significant cost to fix it because it’s a limit of the underlying technology … there IS NO FIX except adding new radios (but you’re limited on frequency availability) or changing the technology (which is the move to WiMAX).

    Time to change your mental model of how the Internet works.

  29. The more you use, the cheaper it gets.

    *Nods* As I suspected, it’s the same as it is in corporate networks. It gets cheaper as you scale your capacity upwards.

  30. PhilK>Isn’t this one of those, “fast, good or cheap – pick two” situations?

    No, because we’re nowhere near the limits of the technology. Think “Moore’s Law” here. It’ll keep getting better, faster, and cheaper for some time to come.

    Let me give you an example. In the Data Center I run, we have the larges switch Extreme Networks makes. The the thing is 100% redundant internally, has a 5.7 Tbps (yes TERA) backplane (or something like that) and can be expanded to something like 57 Tbps (enough to do 192 10Gbps ports switched at line speed). 5 years ago, that would have been unimaginable at any price, 2 years ago that would have cost you > $1M, this thing is < $.25M now.

    Oh, and it runs Linux as does everything that Extreme sells. I've actually seen an engineer drop the console to the shell to check some debugging logs on a BGP issue a while back. Familiar the familiar tools are there, but the end user cannot unlock it (well, not easily anyway :^).

  31. CAPEX – Capital Expendatures. Money spent for capital equipment.

    ESR says: Thanks, that had been my guess.

  32. Oh, and to my “it’s just bits” comment… You can easily substiture “bits” with “porn”. Same thing. :^)

  33. Garrett, extrapolating from your extreme and idiosyncratic peak bandwidth needs to everyone else is silly, particularly when you admit that flat-rate pricing leads to low peak thresholds anyway. Metered pricing would probably lead to much higher peak thresholds as the ISPs would actually be getting paid for those bits, albeit with the overall constraint of peak throughput to the CO. However, as I’ve noted, metered pricing provides more money and incentive to upgrade that throughput.

    Morgan, that’s your “historic” comm-network pricing, 20 years ago? :) Funny how all your comments consist of changing the subject to some irrelevant topic: nobody’s talking about flat-rate pricing with quotas, such as how most people currently pay $30 for 700 cellular voice minutes, which I already said have their place. We’re only talking about flat-rate unlimited plans. Then you point out internet metered pricing in the ’90s, even though I already mentioned that two hours ago in the same comment you quoted. :) Do you even know how to read?

    Don, sorry, going to have to throw the ignorant tag back at you. You do realize that my argument wasn’t based on capex, the “fixed costs” in my earlier comment, but on resource/bandwidth rationing? Oh, that’s right, engineers such as yourself don’t understand such economic issues and only focus on fixed costs. :) btw, you’re talking about data center traffic, whereas we were talking about the last mile, which is composed of “a hard-limited, shared network medium.” Guess what those metered profits pay for? As I already said, expanding capacity through more radios, towers, or fiber. Hilarious how whenever someone makes a grandiose statement like “time to change your mental model of how the Internet works,” they are inevitably the ones who are wrong. :) But the funniest thing in this thread is that esr has never heard the term capex, only the most commonly tossed around word in telecom, and yet has the temerity to call others’ analysis shoddy. :D

  34. @esr> “Back to Apple. Anybody want to bet that Steve Jobs hasn’t already been on the horn to AT&T demanding that they allow the iPhone 4G to ship with tethering enabled? ”

    Why would anyone take that bet when software around tethering and AT&T (specific to AT&T) is already in the 4.0 betas?

    http://www.google.com/search?q=AT%26T+tethering+OS4+beta

    As (nearly) all the articles referenced in the 261,000 results (likely) relate, tethering ships with iPhoneOS 3 (which already runs on every iPhone you can find). The provisioning profile from AT&T disables it. This won’t be an iPhone 4G-specific feature, but it may well be an iPhone OS4 ‘plus”.

    @jay> I still can’t see Apple making a CDMA iPhone. Talk about SKU proliferation. Right now, they get the benefit of making one SKU that will work worldwide, and having only one hardware platform to work with.
    Someone else pointed this out in spades: http://www.appleoutsider.com/2010/05/25/whither-verizon/

    @jay> If they get out of their AT&T exclusivity deal before its 2012 expiration, it’ll be to go with T-Mobile as well as AT&T.
    Except that won’t make any economic sense. T-Mobile has more churn than nearly every other carrier. It would make more sense to screw down their terms with AT&T and continue to rake money there (and demand tethering.)

    @jay> I still don’t think Android’s open-sourcedness is all it’s cracked up to be. What does someone with a Droid have to do to rebuild and reload? Is it even possible?

    no… its not. Not without jailbreaking. And if you’re wililng to go there, you can just run android on a jailbroken iPhone.

    @esr> “Thanks, that had been my guess.”

    anyone who has run anything but the smallest mom-n-pop shop understands the term, Eric. Even googling for CAPEX would have answered the question. (Moreover, Don’s explanation isn’t really correct, but again Google will return plenty of results that explain the term well.)

  35. Ajay — I’ve had unlimited data since my first phone with data, the original sidekick — that was nearly 8 years ago.

    Next you’ll tell us that internet bandwidth went from being unlimited to being metered. ;) Because, you know, in the old days you had to buy hours. X hours per week or month. It was great when you could finally get unlimited bandwidth….

  36. “After last quarter’s unit-sales numbers [that people prefer iPhones] is at best highly dubious,

    Well, let’s see. Chelsea and Amber both prefer to use their personal cellphones (not smartphones) over the (free) work-provided Moto Droid. Chris is using his Verizon Blackberry. The boss whinges about the Droid making pocket calls. That’s because when he finishes a call, the phone puts him into the Call Log tag, and the merest press on any part of the right-hand side of the phone initiates a phone call “and that can be very embarassing” like he might accidentally call the Governor of New York.

    Now, none of these people have iPhones, but … that’s a LOT of resistance to the Droid.

  37. silvermine, thank you for providing the exception that proves the rule. Next you’ll claim that the Sidekick’s unlimited plan is the rule, not the exception. ;)

  38. At one point in time, J2ME (now Java ME) and WAP were the starting points for a discussion on mobile strategy and the web.

    Then, for a brief period of time, one talked about HTML/CSS.

    Now, for a growing majority of mobile strategies that don’t require a global presence on widely varying devices, the discussion begins with iPhone. Smart client is now iPhone app, and in many cases, the app is primary to the experience, not secondary to the browser. And iPad app may soon replace iPhone app as the starting point.

    Frankly, as the adoption rate of iPhone increases and if iPad follows suit, it will become increasingly difficult to argue in favor of a starting point other than iPhone OS. The NPR iPad app, for one, provides a much more pleasant user experience than NPR.org

  39. @Ajay: No, I’m talking about unlimited flat-rate plans. Perhaps I didn’t make it clear that everything went from metered to flat-rate with a cap to unlimited flate-rate. Verizon, the local telco here, offers unlimited all-you-can-eat VOIP phone, including local and long-distance , bundled with their FiOS Internet service and cable TV for around $100 a month. They also offer unlimited all-you-can-eat cellphone service for about $100 a month. Bright House Networks, the local dominant cable company, offers similarly-priced service. I personally have been using unlimited flat-rate everything for years.

  40. The boss whinges about the Droid making pocket calls. That’s because when he finishes a call, the phone puts him into the Call Log tag, and the merest press on any part of the right-hand side of the phone initiates a phone call “and that can be very embarassing” like he might accidentally call the Governor of New York.

    Tell your boss that what he needs to do is press the power button at the top right (when looking at the phone) before putting it in his pocket.It means that he’s got to “unlock” the phone (the slide to the right thing if it’s unsecured or his security sequence if it is) before it’ll actually phone people or do anything else. This isn’t exactly a “Android” issue. The same thing would potentially happen with an Iphone. Hell i made a bag call once in much the same way with a Nokia 6600 (which had physical keys no less).

    android on a jailbroken iPhone.

    That would be amusing to see. Of course it’s probably gonna be a lot of effort to make drivers for all of iPhones soopa secret hardware that is allegedly worth so many patents. But i’d still be amused to see an Iphone with the android home screen on it. It would be like seeing linux and Xeyes running on a playstation 3.

  41. As soon as the monetization model for online video is figured out and tons of HD video starts flowing, they will either switch back to metered pricing or go out of business, with the metered-pricing ISPs taking over their networks.

    So speaking from Australia which is probably the living epitome of what you’re talking about, the smart players went a different route.

    At internode it’s called “shaped” plans. You never get charged for over usage. However as soon as you go past your “plan data” you go down to modem speeds (yeah, 1500kbps to 64kbps is really jarring). Now everyones needs are (kind of) served. The customer doesn’t get hit by surprise nickel and dimes because they didn’t secure their wireless against the local wardriver who devoured their sweet sweet bandwidth and the ISP gets to restrict the amount of excess data that people can download in excess of their alloted amount.

    Other systems have been trialed (Optus Netstats being probably the most unique) but i think this system won out for being simple and relatively hard to game in a meaningful way.

    As an aside, to a certain extent mobile phone plans are somewhat sheltered from this (so far). Given that the amount of space available to leech down hard data is somewhat limited on a phone, it’s harder to push a phone plan as hard as a dsl plan without tethering it to a laptop. (If I were doing planning for a telco I wouldn’t rely on the effect but it’s still there)

  42. Ajay: Your original argument was that carriers would have problems with “network capacity”. If you meant “tower capacity” then you should have said that as the network and the tower are two different things (kinda like the network and the server are different things in a LAN). Carriers address tower capacity issues by using microcells and femtocells. If there are sufficient users in an area to warrant the beefing up of the network, then there are sufficient users in an area to justify that beefiness.

    Also, as the bandwidth of the network increases at the towers, assuming that the backend has been built correctly, the duty cycle of the customers’ usage goes down, freeing up slots for other users and allowing the carrier to shrink the slot size. The carrying capacity increase of WiMAX is a little more than proportionally greater than GSM/Edge or GSM/3G because of this reduced duty cycle and shrinking slots.

    Again, this is largely a CAPEX problem from the carriers point of view. The only time that network congestion becomes a real problem is when the spectrum becomes so crammed that you can’t do anything with it. This happens in large down-town settings line Manhattan, San Francisco or Chicago, but in 99% of the country this is not an issue and likely never will be. If FCC ever gets off their asses on the White Space and recycled VHF TV channels, then this problem will be solved for at LEAST a decade (and that CAPEX number will drop drastically too because of the penetration quality of VHF signals); that’s why Google has pushed so damned hard on this.

    The cost of bandwidth to the carrier is WAY less than linear. I do understand the costs of business. I’ve started several and fought this particular dragon in person.

    You’ve swallowed the marketing lines of the carriers hook, line and sinker. I actually KNOW what I’m talking about, and you’re clueless.

    To be blunt about it, you’re _WRONG_. Deal with it.

  43. JonB>That would be amusing to see.

    Here you go.

    http://www.youtube.com/watch?v=5yO2KQHkt4A

    And yes, there’s lots of stuff broken still, but it sounds like a really good use of older iPhones. Could turn them into Android-based book readers similar to Kindle (just need a really cheap GSM connection).

  44. Amazingly, no one took Ajay up on *this* item:

    > Umm, haven’t phones always had common operating systems, so what changed now?

    No, cell phone “operating systems” (they weren’t that, really, until about 2 generations back) were almost always sui generis; the first really common cross-line OS was probably RIM’s, followed reasonably promptly by several Linux distros, and then Android and the current generation, to a zeroth approximation of accuracy in sequence.

  45. Ajay — well, my exception-proving rule delusion is quite lovely, thank you. ;)

    I’m from a science background. If we just ignored everything that didn’t fit our hypothesis, calling it the exception that proves the rule, they’d kick us out. (Well, unless we were studying global warming. Then it’s rewarded.)

  46. Amazingly, no one took Ajay up on *this* item

    Sorry. My brain can only handle so much garbage data before exceptions start kicking in, resulting in my underlying processes sending such garbage data to /dev/null. ;)

  47. Morgan, if you were in fact talking about unlimited flat-rate (UFR) plans, that fits in perfectly with the analysis I gave before, about ISPs being willing to provide UFR plans for services that have become a negligible portion of their throughput, such as email, web pages, or voice. Do you actually have UFR cellular data? Because that’s the service we were actually talking about. The reason the analogy to historic communications services like telegrams or voice calls breaks down is because those were single-purpose networks, so that as network capacity grew (particularly as most baby bells switched to digital backends), that single purpose became a negligible fraction of network traffic. Wireline and cellular carriers only provide UFR plans on texting or voice because consumers don’t know that they are being ridiculously overcharged for those services today, so they continue paying past high prices even with UFR plans, and because text and voice aren’t going to magically take up a lot more bandwidth, even with the advent of wideband audio.

    However, the internet is a general-purpose network, so we will keep coming up with new ways to use it, that will always put strains on the network. First it’ll be 480p video, then 720p videos and 480p videoconferencing, then 3D videos, and so on far into the future. The wireline ISPs essentially made a bet with UFR a decade ago that the next bandwidth hog, video, would not appear in time to screw them. They have been right so far, though p2p and the rise of low-res flash video sites is making them feel the pinch. As soon as the online video monetization model is figured out and really takes off, they will all switch to metered en masse or go under. The UFR bet is similar to all the dumb banks that bet that housing prices would never fall. It took a decade for them to lose that bet big-time, but it inevitably happened, :) just like it will to ISPs that use UFR. Thankfully the cellular networks still don’t use UFR much and hopefully they won’t make that dumb bet also. ;)

    JonB, how is Australia the “living epitome” of what I’m talking about, are internet video and other bandwidth-hogging services much more prevalent there? If not, your statement is meaningless. As for the soft quota system you detail, I already mentioned a similar system that Sprint supposedly uses earlier in this same thread. You’re right that cellular data usage and capacity aren’t that high yet, but smartphones downloading videos from vimeo are increasing the former.

    dgreer, I never used the words “network capacity” originally, so I don’t know where you get that. A tower is a part of the network, in that it’s an intermediary link and never the destination, unlike a server. To my knowledge, there is no meaningful deployment of microcells or femtocells anywhere today, hilarious that you think there is. “If there are sufficient users in an area to warrant the beefing up of the network, then there are sufficient users in an area to justify that beefiness” – thank you for that tautology, or is it merely a meaningless rewording? ;) You simply assume that capex for tower capacity will increase, whereas one of my main points is that metered plans actually provide carriers with an economic incentive to invest in those upgrades. Funny how you say the only problem is spectrum saturation, right after you pointed out the bottleneck of tower capacity.

    This isn’t purely a metropolitan issue: it’s an issue anywhere there is less throughput to the last mile, because of fewer cellular towers or under-provisioned wireline COs, than the demand for data services. That could be Milwaukee, it could be Mobile. I agree that the fundamental problem behind overall capacity is how the govt has been bought off to not make more spectrum available or easily transferable. Thank you for repeating the familiar argument that fixed bandwidth costs are sublinear, now try addressing my main argument about resource contention at any particular level of network capacity. Please tell me what carrier is marketing metered plans as the future, I’ve never heard of any: it would be nice to find out about a telco that actually understood these economic issues. ;) Hilarious how whenever someone makes a call to authority like “I actually KNOW what I’m talking about,” they are inevitably laughably wrong. :D

    Baylink, hilarious how you don’t even mention Symbian, which has dominated smartphone market share since the inception of the category.

    silvermine, I didn’t ignore your example, I said it was an irrelevant exception. Perhaps you could provide a “scientific” rationale for why it isn’t, that should be funny to read. ;)

  48. how is Australia the “living epitome” of what I’m talking about, are internet video and other bandwidth-hogging services much more prevalent there?

    Your argument, as I understand it, is that flat rate pricing will always be replaced by metered pricing because extreme users will consume all available bandwidth such that the amount they spend towards the maintenance of the network has no relation to the cost of upgrading that network to handle their load. In a way you’re proposing a kind of “tragedy of the commons”.

    Australia is a living epitome(perhaps example is a better word but epitome felt right at the time) for two reasons.

    Firstly, Australia has seen exactly the type of effect I believe you’re describing. There was a network(can’t think of the name of the top of my head) that offered free unmetered transfer to anyone within that network. Within a year it collapsed due to a huge amount of P2P traffic consuming all its bandwidth. Under the framework of ESR’s Iron Laws i’d suggest it’s a special exception where they didn’t make back enough of the costs of the initial buy-in to be able to afford the lesser rate of continued expansion. However thats just a guess as I wasn’t personally involved.

    Secondly, Australia has some interesting properties that make the cost of expanding a network a little more expensive than in the US. For starters we have a very spread out population. Our inner cities have the population densities of outlying US areas. This means that the cost to reach the same number of potential customers is much higher. Also, we don’t just have to reach the local backbone, we have to reach the US (in reality we have to pay people who maintain a link to the US, but it’s still an extra expense). So the breakpoint where consumption will swamp your financial ability to provision capacity will occur earlier in AU than in the US.

    Ultimately what i’m trying to show is that I have seen the situation you’re talking about, and the result isn’t necessarily metered pricing but shaped plans.

    Note that i’m specifically talking home internet here, to my knowledge most Business internet in AU _is_ metered. And really thats not necessarily an unnatural position. The lions share of the drive for expansion is off the back of business use with home use riding it’s coat tails. Perhaps the biggest difference between “normal” and “mobile” internet in this regard is that there is a higher likelihood of people actively using their mobiles during business peak time.

    you don’t even mention Symbian, which has dominated smartphone market share since the inception of the category.

    You might want to look at slightly more recent numbers on that score. If the graph that ESR quoted about 4 posts ago (“Android Rising”) can be believed, Symbian is a bit player these days having been beaten out by Blackberry, Android, IPhone and even Windows Mobile. The graph in question doesn’t have a category for Symbian, lumping it in with “Other” which only beats out PalmOS. So even assuming the only “Other” is Symbian (which is not an unfair assumption) it only beats PalmOS on sales for Q1 2010.

  49. Do you actually have UFR cellular data?

    Yep.

    And what exactly is the difference between a single-purpose and a multi-purpose network? Essentially nothing. Do you actually know anything about the TCP/IP and/or OSI network stacks? Layer 2 and 3 are pretty application agnostic, making them, in essence, single-purpose networks. Routers don’t have to know the difference between HTTP traffic and SMTP traffic, for instance. Packets are packets, bits are bits. New uses are irrelevant.

    BTW–voice networks can and are used for things other than voice. Does that make them multi-purpose networks?

    Ajay, you’re just _WRONG_. Admit it. Listen to dgreer, he knows what he’s talking about.

  50. Ajay> dropped calls and other problems on the resulting overloaded network.

    Later…

    Ajay> I never used the words “network capacity” originally.

    Oh, sorry, I didn’t know that an “overloaded network” wasn’t a “network capacity” issue. My bad ;^).

    Ajay> To my knowledge…

    And there’s the problem.

  51. Ajay:

    how the govt has been bought off to not make more spectrum available

    boggle … If I didn’t think you were completely clueless before this, I would definitely after. Merciful $DEITY, you have no concept of how spectrum is actually allocated, do you?

    1. Jay, while Ajay is an overconfident ignoramus I have to say that “govt has been bought off to not make more spectrum available” is not a completely insane interpretation of the shenanigans around freeing up no-license-required spectrum for UWB and mesh. The telecomms companies really are opposing that tooth and nail for the very excellent reason that iot has the potential to break their business model wide open. Ajay may be contingently wrong about the effects of spectrum auctioning and whatever payoffs are going on under the table, but it’s not essential crazy wrong.

  52. JonB, yes, I think you get my argument, though resource depletion isn’t necessarily caused by extreme users: it’s going to happen because the average user is about to make a quantum jump in data usage, followed by many more in the year to come. And the argument isn’t based mostly on upgrade or maintenance costs but on determining who gets priority for those finite resources, which is what price is always used for. The shaped plans that you describe are basically fixed-rate with quotas, as 64 kbps after you’ve exceeded the quota is essentially worthless as you noted. I think Australian consumers would be better served by metered plans, but if they’re a small part of the mostly metered business traffic, I can see why they might be ignored. As for Symbian, not sure why they’re not in that unit sales graph, but I’ve previously pointed out that they’re still the market share leader.

    Morgan, “Layer 2 and 3 are pretty application agnostic, making them, in essence, single-purpose networks,” hahaha, you’ve taken your nonsensical pronouncements to another level with that one. So what is the single purpose of TCP/IP, pray tell? :D The fact that voice networks can be overlayed with all kinds of data today is completely irrelevant to the fact that historically that was not really possible. This is the equivalent of claiming that because one could theoretically have sent jpeg-encoded images over telegram, it had all the capabilities of the current internet, ie a statement so breaktakingly silly that one has to just shake one’s head. I kinda feel like I’m making fun of a retarded kid at this point, so I think I’ll stop engaging with you and your constant malapropisms. :)

    dgreer, if you had any evidence or argument to make beyond such word-lawyering trivialities, you’d make it, so clearly you have nothing. :D

    Jay, let me quote you your close friend, “Theodore Vail and the old Bell System bribed the Feds to steal the last mile out from under the public’s nose between 1878 and 1920; the wireline telcos have been squatting on that unnatural monopoly ever since as if they actually had some legitimate property right to it. But the telcos’ crimes aren’t merely historical. They have repeatedly bargained for the right to exclude competitors from their networks on the grounds that if the regulators would let them do that, they’d be able to generate enough capital to deploy broadband everywhere. That promise has been repeatedly, egregiously broken. Instead, they’ve creamed off that monopoly rent as profit or used it to cross-subsidize competition in businesses with higher rates of return. (Oh, and of course, to bribe legislators and buy regulators.)” So I’m confused, were these statements “not a completely insane interpretation,” “contingently wrong,” or “not essential crazy wrong”? ;) Hilarious the spin esr tries to put on it when I say it. :)

  53. “I agree Apple has to try to go multicarrier as soon as it can get out of its exclusive with AT&T. But I wonder if Jobs is going to find any willing partners.”

    BLINKERS OFF!! Apple is only “mono” carrier in the US, because the US refuses to use the same standards as the majority of Europe and the rest of the world. I don’t care if CDMA is better than GSM, GSM is my only option in the UK – we DO NOT HAVE a CDMA network. GSM works JUST fine, and we have 3 different iPhone carriers (O2, Orange and Vodafone), at least one other iPhone carrier (Tesco Mobile) that piggybacks on to the network of one of those carriers and I can also chose to get a sim card from, oh let me see, T-Mobile, Virgin Mobile, 3 Mobile or about 4 or so other carriers that piggy back on to the networks of one of those listed previously.

    The US is NOT the world. Please open your eyes a little. I’m sick of North Americans complaining about AT&T in the US or Rodgers in Canada. Get a REAL GSM network installed, that can ACTUALLY handle the capacity you require, stop bitching and get over it.

  54. >>Customers seem to prefer [the iPhone].
    >
    >After last quarter’s unit-sales numbers this claim is at best highly dubious….

    No, no they don’t. The figures are dubious. Especially when you figure in that Admob have pretty much debunked that claim with their figures.

    Statistics are only as good as the body collecting them. Statistics are only as accurate as the body collecting them makes them. I would put it to you, a good salesman can put a positive spin on any figures, so long as the samples taken highlight the desired outcome after tweaking.

    ESR says: Link to the “debunking”, please?

  55. JF Sebastian: So, if I read this correctly, iPhone enjoys a 2M phone lead in the US. That’s pretty damned thin considering the huge head start that Apple had. I’m ignoring the world-wide numbers because those are going to reverse rather quickly when all the Chinese Androids start flying around Asia (which I think the will in large numbers by year end).

    Any idea when the data was collected for these numbers?

    Don

    1. >I’m ignoring the world-wide numbers because those are going to reverse rather quickly when all theChinese Androids start flying around Asia (which I think the will in large numbers by year end).

      Yes, that should be interesting. Android’s lack of required licensing fees makes it a no-brainer for those producers and that market.

  56. @JF Sebastian:

    That article references an article at admob that itself references an actual report from admob.

    The report itself is quite interesting. For one thing, you can possibly extrapolate some of the effect of Apple having a single carrier in the US, although I don’t know how hard android was marketed outside the US, or how to control for that or other factors. From the admob article, you can also download prior month reports, so you could possibly figure out quarterly shipments, except that the single other report that I downloaded (January) didn’t look like it broke out data the same way.

    Anyway, there is a gold mine there, waiting for the right analysis, but at first glance, the article doesn’t appear to support (or even address) henderson101’s claims — it is about the total installed base that is visible from the ad network (and for a start, that’s only half of Apple’s phone shipments), not about quarterly shipments.

  57. @Ajay: I’m simply pointing out that the differences between a ‘single-purpose’ and ‘multi-purpose’ network are _irrelevant_. As, I think, dgreer pointed out, bits are bits from an Internet service provider’s point of view. Anything involving ‘overloaded networks’ is a capacity issue.

  58. @Patrick Maupin The point is most def – figures are meaningless when marketing is involved. I’d also go as far to say, just because the iPhone fails in the US market, doesn’t mean Android will succeed world wide. Going by my own personal figures – I see at maximum 3 Android phones a week commuting in to London, working in the IT sector for a Financial company in the centre of the UK’s financial district. i.e. – where all the money is. I see more Blackberries than I can count. I see almost as many iPhones and/or iPod Touches (as you need to look carefully to sometimes tell the difference.) I see people with BOTH Blackberries and iPhones! Till I start seeing Android phones every day, I don’t believe anything significant is happening here.

    1. >I see at maximum 3 Android phones a week commuting in to London, working in the IT sector for a Financial company in the centre of the UK’s financial district. i.e. – where all the money is.

      Just speculating, mind you, but it may be that the iPhone’s being a positional good biases your sample. That is, the people you meet are more likely to carry one as an “I’m a wealthy toff” signal.

  59. I see at maximum 3 Android phones a week commuting in to London, working in the IT sector for a Financial company in the centre of the UK’s financial district.

    I see almost as many iPhones and/or iPod Touches (as you need to look carefully to sometimes tell the difference.)

    And you’ve got it easy when it comes to spotting an iphone. By contrast, Android phones come in all shapes, sizes and types. Without seeing the home screen (which is fairly distinctive in how it lays things out) or seeing some indicative apps (e.g. android market), identifying an android phone requires an encyclopedic knowledge of all the kinds of android phone.

  60. > Just speculating, mind you, but it may be that the iPhone’s being a positional good biases your sample.

    a different speculation is that in London, (and most of the rest of the world), its all GSM, and that without the distraction of CDMA, (and the oddness around frequency use for GSM in the US), Android isn’t ever going to be quite as successful.

    1. >without the distraction of CDMA, (and the oddness around frequency use for GSM in the US), Android isn’t ever going to be quite as successful.

      I can’t see any reason for these to be coupled. My Nexus One is a GSM phone.

  61. > My Nexus One is a GSM phone.

    That won’t really work well (for data) on AT&T’s network. (Did you know?)
    This is due to the high-weirdness around frequencies used for GSM in the US .vs the rest of the world.

    The iPhone won’t work well on T-Mobile’s network in the US, even if you break the SIM-lock.

    (this is part of what the poster above was referring to when he made light of the US way of doing infrastructure. In Europe all the carriers share a (much larger/denser) GSM infrastructure, (and GSM was designed to advantage this) rather than each carrier building out its own infrastructure (as in the US).

    But I think its simple to understand why Android is getting traction in the US, but … not so much elsewhere.

    Verizon Wireless has 92.8 million U.S. subscribers.
    AT&T Mobility has just over 87.0 million subscribers.
    T-Mobile has approximately 34 million cellular customers in the US.

    The iPhone is only available on AT&T in the US. Why is it surprising to you that people who want a phone that looks and acts like an iPhone, but who are either captive to, or choose to remain with Verizon or T-Mobile are selecting Android phones? Especially with Verizon giving the phones away (2 for 1 deals).

    Now throw in the market confusion about Android, the mis-steps by Google in managing which features are available, and even which software will be available and when, and
    it starts to really show just how grotty the Android market is. Russ Nelson and Jay Maynard have pointed out other factors.

  62. > That is, the people you meet are more likely to carry one as an “I’m a wealthy toff” signal.

    My read on ‘commuting into London’ was that he rides the train(s) and/or Underground.

    As such, he is quite a bit more likely to come in contact with the middle class than the (so called) “upper crust”.

    The “wealthy toffs” are all in their BMWs and Mercs, gridlocked in traffic.

  63. (Man, I wish one could edit posts here.)

    Another data point to look at is how successful Android phones are when sold by AT&T (compared to other ‘smartphones’, including iPhone).
    http://www.wireless.att.com/cell-phone-service/cell-phone-sales/promotion/ces.jsp

    There is a version of the Nexsus One that works with AT&T’s networks, but as noted by both myself and others in this blog, (and elsewhere), it wasn’t a runaway success:
    https://www.google.com/phone/choose

    I don’t have data on AT&T+Android sales but someone here might be able to dig some up. My guess is that the iPhone outsells Android phones on AT&T by a (very) wide margin.

    And once you have that data (assuming my guess is correct), your premise must be re-evaluated. Especially in light of whats happening with iPhone .vs Android in the rest of the world.

  64. Another data point to look at is how successful Android phones are when sold by AT&T

    AT&T’s Android phones are awful and deliberately crippled: http://blogs.computerworld.com/15715/at_ts_backflip_android_phone_is_a_face_plant . Whether that’s AT&T trying to make the iPhone look good by comparison or just being their usual control freaks is hard to say.

    There is a version of the Nexsus One that works with AT&T’s networks, but as noted by both myself and others in this blog, (and elsewhere), it wasn’t a runaway success:

    Not surprising, since you pay the full unsubsidized price and then don’t get a discount on the monthly rate, as you do with T-Mobile.

  65. Hm. I’ve still got a first-gen iPhone, and been wondering about Android…might have to look into that and see how well it works.

    Well, after much headdesking, it’s running. Now, to figure out what it’s good for. The useful starting point is this article at PCWorld, though it’s a bit of a moving target. The one thing that the PCWorld article gets wrong is that there is one step that, as far as I can determine, must be accomplished on Ubuntu Linux, because I was totally unable to get useful drivers for my 2G iPhone any other way: you must use the suplied dripwn binary, and you must run it on Ubuntu. I could not get it to build on any other Linux, and the OS X version failed to produce a usable driver.

  66. Jay Maynard: Congratulations. It’s a beautiful thing to see a nerd in his native habitat ;^).

  67. When we were an agrarian nation, all cars were trucks. But as people moved more towards urban centers, people started to get into cars. I think PCs are going to be like trucks. Less people will need them. And this is going to make some people uneasy. — Steve Jobs

  68. Here in Australia all the carriers offer tethering on the iPhone 3gs, and I think only one did not offer it from the beginning. I wonder how many more iPhones would have been sold in the states if it was the same over there.

    As for features, surely profit motives and product road maps cause companies to try to offer the minimum feature set for the maximum price.

    1. >As for features, surely profit motives and product road maps cause companies to try to offer the minimum feature set for the maximum price.

      Indeed. The problem is that they have competitors who can usually only gain marketshare by offering a better deal. In commoditized or near-commoditized markets (which, despite the carriers’ struggles, is what the smartphone market is sliding towards) this starts a competitive arms race towards the maximum features and minimum price that any player can offer and still make a profit.

Leave a Reply to JonB Cancel reply

Your email address will not be published. Required fields are marked *