Just how big a deal is this? At the very least, we're talking about hundreds of millions of dollars in potential losses to consumers. Although federal regulators haven't disclosed how much they're seeking in damages from AT&T, we can do some math to put a rough dollar value on AT&T's throttling practices. I asked a number of economists, antitrust lawyers and former FTC officials familiar with the process of calculating damages to help give a rough idea of the money that may be at stake here.
First, here's the quick summary. AT&T may have lost consumers anywhere from $300 million to over $1 billion or more.
That $300 million figure may not sound like much, considering the company allegedly misled 3.5 million customers a total of 25 million times over the course of three years. Indeed, $300 million is just a fraction of AT&T's annual revenue — two-tenths of a percent, to be exact. But what's pocket change to a wireless company is big money to consumers and for the FTC: $300 million is 13 times greater than the biggest fine the FTC has ever levied and (for a more apples-to-apples comparison) nearly four times greater than the agency's biggest restitution award against a wireless carrier.
And remember that $300 million is just a conservative baseline figure. Economists and lawyers from both sides are going to argue that damages should be calculated in all kinds of other ways, which we'll go into below.
How do regulators come up with damage figures in these kinds of cases? There are two ways to approach the problem, according to Kenneth Davidson, a former FTC attorney who spent 27 years working on agency investigations and enforcement actions from 1978 to 2005. The simpler, more limited way involves repaying customers who were effectively overcharged by AT&T. These customers paid for unlimited data for a month, but only got unlimited data for part of a month, so whatever fraction of the month that they paid for but didn't get unlimited data, they get back as cash -- refunds for services not rendered.
In exceptional cases, regulators may seek additional restitution. This second way is more expansive in scope than the first, because it does more than simply refund to consumers what they initially paid for the shoddy service. In this case, it might be an estimate of the damage AT&T caused when it kept people from using the data they would have used were it not for the throttling policy. The calculation might also include the value of any business or entertainment opportunities that consumers missed out on because they couldn't get online at the expected speed.
Davidson believes the FTC would be justified in taking the second approach. But, he said, "at the very least, the customers should get a refund for the days they didn't get the service. And that is the simplest mathematical calculation you can do."
So let's just tackle the first approach. Before we begin, you may want to familiarize yourself with the key numbers we'll be using:
The FTC alleges that AT&T's throttling affected 3.5 million unique users.
From October 2011 to now, the FTC complaint says, AT&T throttled those customers a total of 25 million times. Another way to put it: There were 25 million monthly billing cycles for which AT&T allegedly provided, for a portion of the month, slower service than what it advertised. What we don't know is how evenly distributed the impact was; it could have been that a handful of really heavy data users triggered the throttling more often and helped drive the figures up.
The FTC has given us a sense for the average impact of the throttling spread across all users and all instances of the practice. On average, the agency claims, the slowdowns kicked in 18 days into each month, leaving people with 12 days of slowed Internet per month.
When the throttling program began, AT&T's unlimited data subscription was priced at $30 per month, or roughly $1 a day. These prices have remained constant, even through 2014.
Finally, AT&T tested various versions of its throttling program as far back as 2011 but didn't roll out a nationwide, uniform throttling policy until March 2012. For consistency's sake, these calculations assume the 2012 program as a baseline (thus leaving out the five months from October 2011 to March 2012).
With that said, let's start running some numbers. We can begin with a simplified equation that multiplies the number of times the throttling allegedly happened (25 million) by the amount of money AT&T extracted from affected users during throttled periods ($1 per day, or $12 every month). This should yield the amount of money that AT&T collected in subscription fees, from everybody on an affected unlimited plan, just during the throttled periods.
(25 million instances of throttling) x ($12, for 12 days of throttled service per month) = $300 million
That's $300 million in subscription money that customers paid to AT&T, expecting to get one thing but getting something very different in return.
It would be a big deal if the FTC ever managed to get all of that money refunded. But the damages may not end there. Because the 12 days of restricted service represents a per-month average covering all subscribers and instances of throttling, Davidson proposes restarting the calculation — this time, from the perspective of the consumer.
"To calculate the direct monetary harm," said Davidson, "you take the 3.5 million subscribers who were cheated out of 12 days' service each month and multiply that by the $1 they paid for each day of service. You find they paid $42 million each month for services that were throttled. That $42 million cheating went on for 31 months, meaning that the deception cost the subscribers a total of $1.3 billion."
That's a tremendous figure. What this assumes is that the FTC's 12-day average for services allegedly not rendered accurately represents the big picture. In other words, the way it's being used here suggests that every affected subscriber was effectively throttled for 12 days a month, every month, for the lifetime of the program. The way that averages work, this could be true. But without knowing more about how the FTC got its numbers, it's hard to say — certainly compared to the first calculation we ran.
There are other ways of estimating damages. One that the FTC is likely to focus on is how many people were unfairly charged early termination fees for trying to leave their throttled plans. The argument would go something like this: When AT&T began slowing down their users' Internet in violation of their advertising, that amounted to an unfair change in contract terms, and any penalties AT&T charged were ill-gotten and should be refunded. References to early termination fees appear a number of times in the FTC's complaint.
Another option would be to treat what unlimited customers were actually getting as though it were a metered or capped plan and compare that to other plans on the market. If regulators found that customers could've gotten a better deal elsewhere, economists say, then you could award them the difference. Of course, this approach might be thwarted by the fact that unlimited data plans are already a fantastic deal for the consumer, throttling or no — which is how this all came up in the first place. Carriers have an economic incentive to transition people off of unlimited plans and onto more lucrative metered plans, as Ed Felten, a former FTC official who's now at Princeton University, has explained.
It's not hard to see how messy calculating damages can get. As the case goes to trial, expect methodology to become a major question. AT&T is already signaling that it wants to challenge the FTC's underlying assumptions, including, perhaps, the idea that its subscribers were not getting the service they paid for.
“Our network management program was created for the sole purpose of ensuring all of our customers receive high-quality service," an AT&T spokesperson said in a statement. "Anyone who suggests that it is in place for economic reasons is overlooking the fact that wireless capacity is a finite resource that has to be properly managed in order to provide the best customer experience possible, particularly given the ever-increasing data demands of smartphone users.”
The amount of consumer harm may be a different number from what the FTC asks for in court, which may be different again from what, if anything, it wins from a judge. But from where consumers sit, it seems they may be out at least $300 million. That's a pretty safe number to start with.