Why AT&T Should Dump the iPhone's Unlimited Data Plan

Pickens writes "Farhad Manjoo has a provocative story at Slate asserting that while the iPhone has prompted millions of people to join AT&T, it has also hurt the company's image because all of those customers use their phones too much, and AT&T's network is getting crushed by the demand. The typical smartphone customer consumes about 40 to 80 megabytes of wireless capacity a month, while the typical iPhone customer uses 400 MB a month. As more people sign up, local cell towers get more congested, and your own phone performs worse. He says the problem is that a customer who uses 1 MB a month pays the same amount as someone who uses 1,000 MB, and the solution is tiered pricing. 'Of course, users would cry bloody murder at first,' writes Manjoo. 'I'd call on AT&T to create automatic tiers — everyone would start out on the $10/100 MB plan each month, and your price would go up automatically as your usage passes each 100 MB tier.' He says the key to implementing the policy is transparency, and that the iPhone should have an indicator like the battery bar that changes color as you pass each monthly tier. 'Some iPhone fans will argue that metered pricing would kill the magic of Apple's phone — that sense of liberation one feels at being able to access the Internet from anywhere, at any time. The trouble is, for many of us, AT&T's overcrowded network has already killed that sense, and now our usual dealings with Apple's phone are tinged with annoyance.'"

Comments

What uses so much data?

[schnikies79]

I've had an iphone since June. Total data received is just a under 1gb, data sent is around 80mb.

Re:Invest

[Sponge Bath]

...extra money from the new fees rolls in

That is the cruel truth of how ATT, and most carriers, operate. They won't lower the fees for light users, they will only add fees for heavy users. In fact, they would probably add a new fee to light users called "bandwidth usage monitoring recovery fee" to compensate themselves even more for the capability they already have. This fee will, of course, not be counted in the total fee for the remaining 23 months of a contract, just like they do with all the other bogus fees they try to mislabel as some generic 'government tax' like fee.

Fee Fi Fo Fum, ATT can kiss my bum.

Re:Invest

[tsm_sf]

No he's dead on. I just signed up for a new phone and noticed similar language. If you read the fine print you see that it's basically an unspecified portion of your bill that's been given a name designed to make you think you're paying more in taxes than you are.

Another failure to evaluate relevant concerns

[cybereal]

Look, I'm all for consumer fairness. It would be nice to get better prices. But the fact is, whether you have a 5GB plan or an iPhone unspecified/unlimited plan, your averages are still well within the range of limits experienced by both parties. It hardly makes a difference.

The article is basically making the argument that somehow iPhone users should be punished because they're actually using the service AT&T has been selling everyone for a long time. This is pretty asinine. The real issue here is entirely different and entirely AT&T's prerogative. Let me enlighten you:

AT&T's "3G" network, which is actually 3.5G, HSPA... is on the tail end of its lifespan. The technology in all of these handsets depends on it, of course, but it's done. It's over. There is only one last stage of improvement to GSM tech and it's a stretch as it is. Why would AT&T want to invest in expansion of a dead infrastructure? They don't. They aren't going to any more than they have to. They will expand to the last stage of 3G in the largest markets just as they prepare to roll out the same LTE based networks that every other carrier is supporting.

That said, there's no reason to think bandwidth consumption is the primary concern here. The primary concern is one of density. The number of users each relying on the same cell is too great. It's not a matter of how much data they are transferring on that cell so much as that there must be more cells, or cells must be able to handle more concurrent users. That's just a factor of the proliferation of cellular phones and devices. You can't blame the iPhone for this. It's a problem that would occur eventually anyway as the trend towards data enabled devices existed before anyone even knew about the iPhone. Maybe the iPhone accelerated it, but that is no reason to punish people who like a good user experience.

Of course, there's another concern not addressed and that is the exact same concern that effects cable internet subscribers. Cable internet actually works in a very similar fashion to cellular internet. In the case of cable modems, customers share a download node that has a set maximum bandwidth with its uplink. You are sold rates like 12mbps but there is only a maximum of 60mbps at each node. So if more than 5 people all try to use 12mbps at once you won't get what is promised. However, because most people don't use nearly the maximum pretty much.... ever... the cable companies overprovision the network. They get away with it because the statistics generally match up. However, if you're unlucky enough to live in a neighborhood full of download happy geeks, you're going to hate your internet connection.

The same issue exists in cell towers. A give GSM cell can handle a fixed maximum number of communication slots each functioning as a statically wide band of communication. When a device ramps up from basic voice to data, to higher speed data, it will consume more slots. Or it won't, if there are none available and it will just stay slow or not connect to data, or whatever. So basically if you have 1000 slots on a given tower, and full 7.2mbps hsdpa+ requires 12 of those slots, you can see that there's a fixed number of people who can possibly access the network at full speed. Add to this the already common problem of the actual backing internet connection experiencing the exact same kind of limitation and you can see that infrastructure is a problem of density, not of actual transfer totals.

So, the lesson here is that more uplinks are needed so that uplinks are not as central a point of failure as they are today. What you'll earn is that cells are relatively evenly distributed across all markets but not all markets have an evenly distributed level of usage from consumers. People in metro areas will note the worst performance because there's simply too many people in one place. You'll note the epic failure of networks during large technical conventions with a 1000+ simultaneous attempts at liveblogging the latest