Attendance as a Predictive Metric

Ethan over at Actionless Activity does an attendance analysis of the 2009 WNBA Regular Season.

His idea: look at attendance numbers and see if they indicate anything about the team: he sets a mark that if a team claims an average of 7,000 fans per game, it should be profitable and if it claims an average of 10,000 fans per game it is probably a "legitimate" franchise with a long term future in the WNBA.

However, given the average attendance figures for the year, you would have to conclude that every team is profitable except for one (*) - the Connecticut Sun - and two teams, the Sparks and Mystics would be "legitimate".

The problem with the analysis is that attendance numbers...really don't say much of anything. First, every league inflates its attendance - they all do it, not just the WNBA. Even football has been known to give away tickets to move crowds which are close to selling out up to official "sellout" status. I've seen "crowds" of 16,000 at pro baseball games that looked like if the crowd rushed the field to attack the players I'd give the players even money on winning the fight.

In some places in the WNBA - like Connecticut - attendance figures are roughly accurate. (Connecticut doesn't have much incentive to lie.) In other places, like Detroit, the attendance figures might as well be drawn up out of hat. Attendance figures can indicate how confident the team is in the product, or...the lack of confidence in the product: "hey, if we blow up these figures to 7,000 maybe someone will take notice!"

Part of the problem with attendance is that there is no official "turnstile attendance" - how many actual fannies are in seats during any one game. If those numbers are kept, we'll never see them. (No league will ever show them - do you really want people to know the difference between claimed attendance and physical attendance?)

Another problem is what do you do about season ticket holders that miss games? The STHs have paid for all of their seats. So why wouldn't you count them in attendance? If someone has paid for a seat but doesn't show up to take it, the franchise has their money and it's almost as good as someone attending. Maybe those people would have attended if something hadn't kept them from coming - or maybe they wanted to support the team without the labor of showing up for games.

One poster at RebKell stated that one should not assume that once attendance reaches a certain point, profitability is implied. (Except if you have 250,000/game.) One must factor in the cost of:

* arena rental: if you also own the arena (Connecticut), you need fewer attendees to remain profitable

* who controls parking rights? concession rights? For example, Atlanta doesn't get any concession rights - if you buy a hotdog at Philips Arena (**) not a single bite of that dog goes to the Atlanta Dream. If teams get a cut from other sources, it lowers the number of tickets you have to sell to remain profitable.

My understanding is that Los Angeles has something called "premium seating" - if you're a ticket holder of a premium seat (for the Lakers, for example), all of that money goes into a pool and each of the arena franchises gets a slice of that cash. The Sparks get only a tiny sliver of Jack Nicholson's money - but that tiny sliver is enough.

My conclusion: looking at attendance is, at best, an imperfect metric. What might be more interesting to look at is change in attendance as a metric. For example, in 2008, the Comets claimed attendance dropped by 1,581. They had moved to a new arena, of course, but their team high in their new digs (7,261) wasn't even as high as their claimed average from the year before (8,166). That should have been a sign of trouble right there.

____


(*) - I stole this joke from a story about the Dallas Cowboys.

Some lout watches the owner of the Connecticut Sun walk down the street. "Hey, dumbass!" he shouts. "I hope you like losing a million a year on a WNBA team!"

The man takes off his hat. "I guess I better do something, then, or I'll be bankrupt in a thousand years."

(**) - I prefer bratwurst, myself.