Eduardas Valaitis
joined Pricewaterhouse Cooper’s national economics and statistics practice as a consultant in 2008. He earned his PhD in mathematical statistics from Yale University in 2005 and taught as a tenure-track assistant professor in the mathematics and statistics department at American University.
Increasing amounts of data become available to businesses every day, and we, statisticians, are uniquely qualified to help them prepare, organize, and analyze these data. The term “business intelligence” is often used to describe such analyses, and most industry leaders rely on statistical analyses to guide business transformation and growth in an informed, thorough, and actionable manner. Therefore, new graduates with master’s or doctorate degrees in statistics are highly sought after; however, to land an industry job, statistical expertise alone is not sufficient. The following are some tips that will help both job-seekers and those who have just started their career as industry statisticians.
Avoid Narrow Specialization
Leading statistics programs in the United States are generally focused on providing their advanced-degree students with solid theoretical foundations and the ability to perform exceptional research in a narrow field of study. However, as a statistician outside of academia, you will mostly use established and credible methods to analyze large amounts of data in a compressed timeline. These methods will likely span a number of fields, such as time series, clustering, or sampling. As such, those “applied” courses that so many of us looked down upon in our days as graduate students (i.e., data analysis, SAS programming, sampling methods) are useful for building analytical skills industry statisticians rely on to succeed. Moreover, while your dissertation and studies may have been focused on discrete noninformative priors, it would help tremendously if you are proficient in a number of analytic areas.
Improve Your Software Proficiency and Programming Skills
Industry-standard statistical software packages such as SAS, SPSS, and EViews are becoming increasingly user-friendly; however, they often provide point-and-click interfaces that do not require the user to be proficient with the underlying programming language. I strongly recommend you avoid relying on the point-and-click approach to data cleaning, organization, and analysis. Instead, you should spend time learning a few programming languages and, in general, ensure that your programming/logic skills are excellent. As an industry statistician, you may have to learn your employer’s proprietary programming environment, and having experience in a multitude of languages will be useful.
Finally, you should be proficient with such industry-standard software as Excel and Access, as they may be the only software you have at your disposal. Therefore, you should not treat them as mere data repositories, but become capable of using their full functionality, including functions, shortcuts, macros, queries, and add-ons.
Become More Than a Technical Asset
While certain industry jobs may rely solely on your technical skills, it is more likely that you will have to develop and rely on your “soft” skills to sustain a long-term successful career. Sample size calculations may be all you are doing as a junior statistician at a biopharmaceutical company, but you may, as you advance, have to explain clinical trial results to an FDA panel and investors or effectively convey the importance of your analyses and results to a CEO that is about to cut your department’s funding. As such, developing your oral and written communication skills is imperative for your success.
To do so, take an English writing class to boost your writing skills, subscribe to The Economist or Bloomberg Businessweek and learn from their written communication style, use opportunities to present your work to nonstatisticians and tailor your presentations and verbal communication skills to be engaging to such an audience, and learn what communication styles are appropriate in different circumstances.
Focus on the Impact of Your Results
While we, statisticians, often get excited about asymptotic properties of a newly minted estimator, your clients/employer will most likely be uninterested in such minutia. In your work, the type of analyses and methods used will be of little interest to your audience; instead, they will want the answer to the “so what” question. Thus, when performing analyses, you should not look at the underlying data as simply numerical or categorical values, but rather spend time trying to understand the business processes that yielded the data and what these data tell you before you proceed with your analyses.
For example, you may be asked to assess health care claims data to devise an algorithm for flagging potentially fraudulent claims. You should familiarize yourself (through research or conversations with specialists) with the way health care claims are processed, common fraud types, and other relevant details. The responsibility of understanding the data and interpreting the results so they are meaningful to your clients/employer lies with you, and you should be proactive in educating yourself because you may be handed the data without much of an explanation.
Be Flexible
Finally, one of the most important factors for landing a job or successfully advancing in your career as an industry statistician is your ability to be flexible. Being flexible may entail the following:
- Being able to consider a number of analytical approaches, instead of relying on the technique you are most comfortable with
- Collaborating with colleagues whose working styles differ from yours
- Being willing to work extra hours or on weekends and travel for a few days to meetings on short notice
- Juggling a number of competing priorities and saying “no” when you have too much on your plate
- Understanding when you need to make a decision on your own or seek the advice of a specialist or your boss
If these all sound like fun aspects of a job, rather than insurmountable challenges, then a job as an industry statistician may be just right for you.
Dr. Valaitis,
Your outline is excellent and timely. Google’s Chief economist Dr. Varian offers similar guidance. Expanding knowledge into the application of statistical decision making to more and more of our daily affairs is a revolution started before Demmings but not fully implemented here in the United States. This is why American Management continues to “chase its tail” with failed attacks on Demmings designed products like the Honda Accord. I have a hard time believing it is a lack of brain power here in the states as proposed by business minds. The American tradition of rational thought backed by science and it’s hand maiden statistics goes back to our inventive founding fathers Jefferson and Franklin and cultural beginnings in Greece. As American’s we are statistical and curious by nature. We are also fearful and wrong sometimes. With a pedigree of statistical geniuses and monumental visionaries like Bill Gates why do we have so many challenges? I believe I have an answer for the young up and coming statisticians. Since 1980 and the Reagan Revolution the US has used the MBA as the primary decision making rubric. The world of business is exciting and as a business owner – I love it!
But with an undergraduate degree in political science and philosophy I like to step back and look at the big picture. Maybe the spinning blue planet we live on is not just built for business. Maybe our data sets ought to expand into the areas of public measures that lie outside the MBA Chartered Financial Analyst perspective. I propose that your generation does this. Document and study patterns in all areas of life. You will see connections between social policy pursuant to human capital development and long term cultural value growth at home and abroad. The number of future markets is experiential once we fully implement the age of “enlighten decision” making.
Enlightened? You would think that 30 years of MBA degrees being pumped into the US Economy would have resulted in a growing standard of living of all Americans. Instead ANOVA suggests a negative relationship between MBAs/capita and Median/Mode of the average American with respect to (health, wealth and education statistics). So why is there a negative relationship between the quantity of MBAs per/capita and the social health indicators of health, wealth and education of the average american. Are American’s really the relatively sickly, low IQ bunch that are too lazy to work as the business community would have us believe. Soci-biology would suggest not since American’s trickle in from all over the globe so any genetic defects would washout.
The reason for this mix-match of a priori knowledge and ANOVA is because our cultural assumptions are failed. Contemporary American decision making is based on private sector variables unfit for public problem analysis. In a sense, we are using the wrong tool and continuing to sharpen this tool in error. For America to continue to focus on the MBA decision model for national planning is like sharpening a tile spade to drive wood screws. Each time you point out its the wrong tool business minds hire another marketing team to wine and dine you so business can continue to invoice adjudicative sharpening fees on a business model from multiple technical generations in the past.
This is why statisticians need to expand there applications outside the MBA box of CFA metrics. MBA’s are a critical “component” of our economy but they are clearly not great at managing it. They simply don’t have the right tools to lead nations.
Godspeed to the next generation of statisticians – you are headed into a world that needs your skills like never before.