Since you asked me …. four tips on getting started in the statistical consulting business

^—– Old advice-giving person

This week I had the pleasure of speaking via WebEx with Rebecca Ottesen’s statistics class at California Polytechnic University (fondly known as Cal Poly by the folks at home). The students sent a number of questions in advance, many of which boiled down to how do you get started in consulting. Since graduation and the job market are looming on the horizon for many people, I am getting these types of questions often lately. Here, in my usual random fashion, are some rambling answers.

1. Decide whether you want to be a consultant. Most bright graduate students are probably encouraged to become professors and most bright undergraduates are encouraged to go to graduate school. I encourage anyone planning on graduate school to read this Chronicle of Higher Education article, “Graduate school is a means to a job” . My total debt when I finished graduate school was zero. That’s right, nothing. Nothing against professors, I was one myself for many years, but your professors telling you that you need to go to graduate school and apply for academic positions is kind of like my realtor friends who tell me how good of an investment a house is. They may both be right, but the advice is not completely disinterested. You’re a lot more likely to get work as a consultant if you have a degree, preferably a Ph.D. That degree doesn’t have to be  in statistics from a mathematics department. In fact, if I was hiring for work I need done, I’d be more inclined to hire someone from an applied field. So … can you get a graduate school to give you a fellowship to attend? If you’re not that level of good, you have less probability of making it as a consultant, so I wouldn’t go into debt to get a degree if I was you. Do you want to go into business instead of academia? That is a very big decision. For me, there were four big incentives for choosing business – my children. I tried paying the kids’ tuition bills with prestige and reprints of my journal articles and the Archdiocese of Los Angeles office informed me that those were not legal tender.

2. Get experience with real data. Download some open source data and do a paper or two for your courses. Write a masters thesis or dissertation. Work on a  research project with one of your professors. Do an internship. I very rarely take on interns for the same reason I hate training new employees – it takes up my time. The less I have to teach you to do when you come to work for me, the more productive and happy I will be. Not only am I paying you to learn, I am losing money in billable hours teaching you. Of course, new people need to be taught the ropes, but the less I have to do, the better. Chief among things you need to know are some ideas on how to beat data into shape before you analyze it. If every data set you have ever worked on is squeaky clean like those examples in statistics textbooks, I really don’t have time to teach you everything you need to know. I’m sure you are very smart and you can learn it – somewhere else.

3. Go work for a big institution. Pick a large consulting company, a big university, a major corporation – and go work there for $50,000 a year and all the shit you can eat, or whatever it is they are paying these days. They won’t pay much but you aren’t worth much. I don’t care where you graduated from or what equations you can spout. You have no experience and you have never really done anything independently. (No disrespect, we all started with no experience.) I don’t need you to prove anyone’s theorem for me. Work hard, listen and learn. You’ll get a chance to work on projects with people with a lot more experience. You’ll see a lot of code written by other people, review reports.

4. Take every chance to learn that you get. If they offer training, jump at the chance, not to impress anyone but to learn as much as you can. File away every good idea they expose you to. Go to REAL training, not some crap on how to use Microsoft Project (do they still make that?) or how to write a business email. Take programming classes, attend conferences. If there are technical people giving talks at your company – go! Present at conferences – explaining what you know and do to other people will help you a lot. You’ll often find that, in explaining your procedures, you realize you could do it better. If you didn’t get a fellowship to graduate school, your company may very well pay for your graduate courses. Don’t whine about not having time to go to school after you worked all day. Are you nuts? Someone is willing to pay money so you can take graduate courses in programming or systems design for free? Do it! And do it at the best school you can find because your purpose is to LEARN not just to get a degree. I have had students tell me, “All I want is a piece of paper.” Well, guess what, clients are looking for someone who has more than a piece of paper.

Notice I didn’t say anything about networking? There are two reasons for that. The first reason is that all of those things I just mentioned – graduate school, big companies, training classes, conferences – are the perfect way to meet people who one day might need your consulting services. The second reason is that unless you do these things you won’t know anything worth consulting on. It doesn’t help if people know you if they also know you are a clueless moron.

On the other hand, if you follow those four avenues, you are very likely a few years from now to have people start asking you,

“Do you do any consulting by any chance?”