Modeling Consumer Behavior

Over at Adaptistration, its Take A Friend to the Orchestra Month (TAFTO). I am not writing this year, but I am participating in a sense. The orchestra will be performing in the theatre I run.

Drew prefaced today’s entry with a promise that it would wow readers with the concepts it was presenting. I have to say it certainly did for me. Bill Harris of Facilitated Systems creates a computer model to test if Drew’s TAFTO program is beneficial for orchestras in comparison with paid advertising.

Now since he is dealing with statistics and computer programs, it isn’t the easiest of reads. On my first read through I absorbed enough to realize it was providing enough valuable insights to read through again a couple hours later. If I understand correctly, one can copy the program he has written and use it in the simulator he suggests to produce results specific to ones organization.

I was intrigued by all this so I followed a link back to Bill’s blog and came across an entry on the Knight Foundation’s Magic of Music Final Report. Not two weeks ago I had cited a portion of the finding of this report to a group and now I see Mr. Harris telling people to be careful about the conclusions they drew from it.

He quote from page 32 of the report-

In trying to profile the factors that might predict a ticket buyer, one statistic stood out: 74 percent of them had played an instrument or sung in a chorus at some time in their lives.

What he says this appears to be saying is,”the probability of someone having played an instrument or sung in a group, given that they were a ticket purchaser, was 0.74.”

But what he says you really want to know is the probability that someone will buy a ticket “given that they played an instrument or sang in a group.” That may be what you assumed the report was saying because you hope that people who play instruments and sing (or perform in a play, paint, etc) will patronize your organization.

My assumption about the findings in the Knight report was that people who had music in their background might be inclined to attend later in life, but I didn’t see a cause and effect relationship. It merely seemed that people with a musical background shared were an affinity group within symphony attendees.

However, under the suspicion that inclination to attend wasn’t any different than cause and effect assumption, I posted a comment to Harris’ latest blog entry asking if I was making an erroneous assumption.

We shall see what he says. In the meantime, the lesson here is to read those statistics with a careful, critical eye.

About Joe Patti

I have been writing Butts in the Seats (BitS) on topics of arts and cultural administration since 2004 (yikes!). Given the ever evolving concerns facing the sector, I have yet to exhaust the available subject matter. In addition to BitS, I am a founding contributor to the ArtsHacker ( website where I focus on topics related to boards, law, governance, policy and practice.

I am also an evangelist for the effort to Build Public Will For Arts and Culture being helmed by Arts Midwest and the Metropolitan Group. (

My most recent role was as Executive Director of the Grand Opera House in Macon, GA.

Among the things I am most proud are having produced an opera in the Hawaiian language and a dance drama about Hawaii's snow goddess Poli'ahu while working as a Theater Manager in Hawaii. Though there are many more highlights than there is space here to list.


1 thought on “Modeling Consumer Behavior”

  1. Joe, thanks for your comments. You seemed to get everything pretty much right. I responded to your comments about the Knight report on my blog; I wanted to point out one other thing here.

    You write, “If I understand correctly, one can copy the program he has written and use it in the simulator he suggests to produce results specific to ones organization.” That’s true: you can download the model, install a copy of the simulator, and run the simulations yourself. You’d need to make up a so-called “simulation file” to tell it what to do, but I could share one to get someone started.

    What I wanted to highlight here is the purpose of such “system dynamics” models. They are almost always intended to explore the ramifications of policies, not to produce “point” estimates. That is, they will tell you that WOM marketing is likely better than advertising for building a community of concert-goers, but they won’t generally tell you (at least with any credibility) how many tickets you’ll sell for Friday night’s concert with the Beethoven sixth symphony and how that might change if you had programmed the third, instead.

    In this sense, policies are the rules or guidelines by which we make decisions.

    What you can do is tailor the parameters of the model to the specific situation in which you find yourself. Given that customization, you might indeed find that, in some situations, you would draw other conclusions about the relative usefulness of WOM and advertising.

    You can also modify the structure of the model, as I did when I added the TAFTO component, to better represent the problem you’re facing in a particular situation.

    So the first key lesson about a system dynamics model is that it’s designed to evaluate policies, not predict the future.

    A second lesson is that you use system dynamics models to address problems, not to model a system. That is, you create a model to answer specific questions (“Is TAFTO any good?”) rather than to model the entire workings of a particular symphony orchestra so you can ask questions later. The former is a doable, useful task. The latter has no stopping point until you’ve replicated the entire orchestra, potentially down to the strings on the instruments and the oil in the valves, for you never know what question you might want to ask.


Leave a Comment