The Atlantic Home

If you're like me, then you've been eagerly awaiting the day that all airlines equip their planes with Wi-Fi. But if you're really like me, then you also don't want to pay much for it. The New York Times yesterday explained that this is a serious problem for airlines wanting to ramp up their in-flight Wi-Fi services. Even though all planes offering Wi-Fi seems inevitable with the vast proliferation of the internet, the cost is holding airlines back. I wonder, though, if the problem could be airlines' pricing decisions.

So just how expensive is in-flight Wi-Fi for airlines to install? Pretty expensive. I spent the past two days chasing down sources for this piece (which in blogger time translates to like a month) to try to get greater detail on those costs. I finally spoke to an industry insider who explained that initial capital costs can easily be in the ballpark of $200,000 per plane.

And then there's the ongoing usage costs that airlines face for their passengers to use the service. That information is so sensitive the industry insider I spoke to wouldn't even provide any broad data on background. But the insider did comment that it was not a terribly significant cost, and through context, I believe it's probably only a few dollars per user. So let's assume it's $4 -- it might be less, but it doubt it's much more.

Let's do a thought experiment. Additionally, I spoke to a major airline spokesperson who provided me with some of their typical plane usage statistics. That will help validate my assumptions. I learned that a jet that holds around 150 people makes around three trips per day. Each jet in their fleet also experiences around 2 weeks of downtime per year. If you crunch some numbers, assuming 80% passenger capacity (per the airline I spoke with), you can conclude that this plane carries around 126,000 fliers per year.* That's how many potential customers Wi-Fi service could appeal to.

The bad news is that current rates of usage are only about 10% to 20%, according to the Times. Using the low point in that range, that means only around 12,600 Wi-Fi users per plane per year. But is this really bad news? Those rates are based on a fee of around $12 a pop. Less the $4 usage fee, that amounts to revenue of around $100,000 per year -- which means it pays for the initial capital investment in just two years.

Of course, this calculus is simplistic and approximate. But it should hold up for most kinds of planes, as smaller planes fly more often and larger planes fly less often so passenger numbers probably don't change much overall. It's also only based on one carrier. So there might be some margin of error from carrier to carrier.

The first thing that occurs to me is that $12 is an awful lot to pay for internet service. I generally refuse to pay that at a hotel -- and that's usually for 24 hours, not a short plane trip. So one thing that seems clear to me: a lot more people would be willing to use the service if it were cheaper. After all, that's how a demand curve generally works.

This leads me to the question: what the shape is of that demand curve? I have a hunch that the airlines might assume the demand curve looks something like this:

But what if it, instead, looked something like this:

In both of these graphics, the pink area is the gross revenue (not including usage) that the airline would obtain by charging $6 for access. As you can see, the revenue is much greater if the demand curve takes the shape of Curve #2, rather than Curve #1.

Airlines would be better off charging less if the curve is flatter. So using these curves along with all our prior assumptions, what if the airline charged $6? With that $4 usage cost assumption, Curve #2 would produce $126,000 per year in net revenue -- more than the $100,000 earned at the $12 price. That also pays for the initial capital expenditure more quickly, in 19 months. For Chart #1 if you charge $6, that revenue drops to $63,000. That's because utilization changes from 25% to 50%, based on those curves' shapes. The shape of the demand curve matters a lot.

So which curve is it? You'd have to construct it by asking people what they would be willing to pay, so that's what I want to do. Vote in the poll below and let us know.

One final note: another possibility that airlines might want to consider is selling access based on time. For example, if I am on a six hour flight, internet access might be worth a lot more to me than if I'm on a one hour flight. As a result, maybe access should be charged in 30 minute increments, for example. That would also allow airlines to utilize more of the demand curve, as if you had to pay $1 for every 30 minutes you use it, then fliers could just use it for a short time but the airline would still get additional revenue. I'd call this the in-flight internet café approach.

In any case, here's that poll. Let us know what you think!

* As an amusing anecdote, I did this estimation out of thin air before speaking to the major airline, just based on what I thought was likely the case. My estimation was somehow exactly what their numbers also provided. So, I'd like to take this moment to thank my college physics classes for cultivating my apparently very potent estimating skills.