Truck and Barter

Outrage! British Airways is selling full price seats on aircraft instead of giving them to stranded passengers. No doubt this secures multiple advantages to the airline. But its effect on passengers must be thought through, and examined empirically. By default, I'm not against the pricing scheme, unless I know how it really affects allocation.

The first question in allocation is, "who should get the seats?" And many people in this situation default to a first-cancelled, first-rebooked standard. However simple, and deceptively fair, that is clearly not an efficient allocation. We want the most desperate home first, no?

The whole point of an allocation system is to measure desperation, and BA's pricing is one way to allocate tickets to those most desperate. I could even say that using sky-high prices to hold some seats open to the highest bidder is remarkably fair.

Yes, BA's pricing does favor those willing and able to pay full fare over those willing but unable to pay full fare. But that doesn't imply those able to pay are less desperate than those unable. And more importantly, it lets people who weren't previously booked and cancelled the opportunity to fly soon if they are truly desperate.

In the end, only by leaving seats for sale at the highest prices, will there be substantial numbers of seats open for those truly desperate enough to wait in the airport on standby. So, by setting high prices for a limited number of seats, the BA scheme winds up discriminating between two types of people who are unable to pay: those willing to wait on standby and those unwilling to wait.

Is there a better way to measure desperation without a price signal?

Now, I should note that, if flights are returning half empty, than BA is absolutely incompetent, and deserves to go bankrupt.

There are a few data points resulting from the unexpected news two weeks ago that Hulu made a profit in 2009.

• $100M in revenue in 2009.


• 903 million videos in January 2010.


• Triple year over year growth in vids served

This implies 300M vids served in Jan 2009, and an average increase in 50M every month through 2009. An average of 600M per month over 12 months is 7.2B videos. $100M/7200M implies 1.3 cents of revenue per video served. If 50% is shared with owners, we're talking about 0.65 cents per vid served. Wow.

How much money should vid producers expect from this model, since it doesn't seem possible to use this Hulu to pay for a substantial portion of the costs of expensive television programming?

Let's see what happens.

This is the Dinar And Discussion Page for April.