Part Three: Interview with Mike Smith – Associate Professor of Information Systems and Marketing, Heinz College, Carnegie Mellon University
This interview began in part one discussing Dr. Smith’s graduate student experience and job search and continued in part two discussing Dr. Smith’s experience and advice as a faculty member. In this session we will discuss his perspectives on research.
Perspectives on Research
What is the “secret” of your productivity?
I’m a firm believer in treating this as a job. Our undergrad days have led us to believe you can work a twenty-hour day and that has to be twice as good as a ten-hour day. At least for me I’m not productive for more than ten hours a day. I think it helps to recognize that I only have ten hours to get a certain amount of work done and I need to intentional about keeping track of what time I am goofing off and what time I am working on the things I should be working on.
How would you describe the process from start to finish of one of your “typical” papers?
I tend to get the ideas that I’m most fond of from the press and the media. I tend to advise my students to look to questions the media is asking, as opposed to the two paragraphs at the end of somebody else’s paper where they talk about future work. You can be sure, that if someone tells you at the end of their paper that “future work should do” this list of things it means one of two things: either some referee asked them to do it but it was too hard, or it is a problem they are already working on.
Most of the topics that I have found most interesting have come from quotes in the industry that say “we think this is happening but we’re not entirely sure,” so start with that. For me it is about finding the right data to ask that question, and then trying to look at that data pretty quickly to see if it has any hope of saying something that is going to be interesting. We’re working right now on a project where we’re thinking about the question “if you take content off a legitimate digital distribution channel does that influence piracy.” The first thing I asked my student to do was take two or three popular pieces of content and just graph what piracy looks like, and if we see a spike the day after something goes off this channel we’ve probably got something we can publish.
I’ve really enjoyed some of the economics papers that start from just the pure descriptive statistics. Austan Goolsbee is extraordinarily good at this, if you read one of his papers it will say “Here is an interesting question, we got some data, we ran the simplest possible test on it to see what’s going on and it sure looks like something is happening. Of course that is a very unsophisticated test so we’re going to need to go and do X, Y, and Z. When we do X we see it, when we do X and Y we see it, and when we do X, Y, and Z we see it.” Building up from the simplest possible test suggests that there is something interesting going on here, and then you bring in the more sophisticated econometrics to tease that out.
I have gotten in the habit of starting to write my papers earlier as opposed to later. Once you found that there is something going on in the data, start writing the paper. I’m not averse to writing an abstract before I write the paper. I think that keeps me focused on what this paper is doing and what I expect to find. You’re obviously going to change that abstract a lot, but having that abstract in mind from day one is a helpful discipline. I frequently tell my students to go write an abstract of what they think this paper is about when they start a project.
All your papers seem use very sophisticated Econometric methods. How did develop your research methods’ skills, and how do you keep up on them?
It’s an important lesson for students to get as tooled up as possible during your course work, those are the skills that are going to carry you through to tenure. You can keep building on those skills but if you don’t have a decent foundation you’re dead. MIT forces you to do this, so I can’t attribute this to anything other than being forced to do it, but I’m glad I took some pretty tough classes because it brought me out with a set of skills that I could then build on. If you don’t have those skills to start there’s no way you can generate them “on the job” while trying to get tenure.
Many of your papers have become influential in policy decisions. Could you discuss the ways your research has been applied in the real world?
Again, the joy of this job is that you get to work on problems you find interesting. I have tended to gravitate towards problems that are talked about in industry journals. We had some influence on policy associated with peer-to-peer networks and that was just because I thought there were a bunch of interesting questions on how these networks should be arranged and organized and we went out and looked at those. I think we’ve had some influence on whether legitimate digital distribution channels cannibalize your physical sales or cannibalize piracy. I started off that paper comparing a quote from Steve Jobs who said, “You’ll never be able to stop piracy, what you have to do is compete with it”, with another person in the industry who said, “You can’t compete with free, that’s an economic paradigm that doesn’t work.” So here we have these two very bright, very prominent people saying completely different things. The question is what really happens in the world and those are the sort of topics I find fun. I also think they end up being more broadly influential.
In addition to your many journal publications, I counted about 5 or 6 pages of conference publications as well. Some professors seem to encourage heavy conference submission/presentations, and others shy away. What is your philosophy on the benefits of conferences, and what advice can you give on how to get the most out of conference attendance/presentations?
There’s obviously a balance. Getting more conference publications is not going to get you tenure, at least in IS, computer science is a little bit different. Going to a conference for me is either (A) there is a paper I want to work on and this is going to force me to work on it or (B) this is an important and influential conference and I want to make sure I’m getting my work in front of important and influential people. I’ve tried to be intentional about cultivating a set of people who are in the IS/Economics community so I go to the Workshop on Information Systems and Economics (WISE) every year, I go to the International Conference on Information Systems (ICIS) every year, and then I’ve tried to go to other disciplinary conferences where there is some intersection. For example conferences like Marketing Science, or the International Industrial Organization Conference where there are people interested in work at the interface with IS who I can get my work in front of. But you need to recognize that going to more conferences, at least in IS, is not going to get you tenure, it’s all about “A” publications.
What research are you working on currently?
Right now we are looking at whether digital distribution channels cannibalize sales in traditional retail channels and whether digital distribution reduces piracy. The conventional wisdom today is that digital channels are going to cannibalize physical sales, but you are never going to recover from pirated products. We’ve tried to look for opportunities to study this phenomenon.
We started by collecting data on piracy, and collecting data on DVD sales at Amazon and waiting for an appropriate event. The event we used was NBC taking all of their content off of iTunes, which offers a great natural experiment. We have all these people who are cheerfully buying on iTunes and now we can see where they go.
The short answer is statistically none of them go buy DVD’s at Amazon.com, and a whole bunch of them go to piracy. There is about an 11.5% increase in piracy which is about twice as large as the number of sales that NBC was getting on iTunes. You can start to play around with a bunch of interesting questions such as why is it the case that consumers seem to choose between legitimate digital purchases and illegitimate digital downloads and not between legitimate digital purchases and legitimate physical purchases. We’re playing around with a bunch of topics that come out of those questions such as what other settings in which this might this be true, and other setting in which might this not be true, and what types of content are you most likely to buy vs. pirate.
Check back for part four of this interview. Dr. Smith will discuss his perspective on teaching and his thoughts on Information Systems as a discipline.