An Interview with the Godfather of Data Analytics, SAS’s Jim Goodnight
Jim Goodnight is one of the great technology entrepreneurs of the past fifty years. His emphasis on data analytics as a business model starting over 40 years ago with the development of Statistical Analysis System (SAS) while he was an academic at North Carolina State was quite prescient, presaging the analytics boom that has taken over so many industries by multiple decades. Over the years, SAS has been used by pharmaceuticals companies to help them analyze their drug pipelines better, by banks to help them assess who to give credit cards to, and to an increasing extent by a wide array of companies to assess fraudulent activity and risk management, which Goodnight suggests will be a significant area of growth for SAS.
Goodnight has built a multi-billion dollar software company without going public or seeking suitors to buy the company. He points to the advantage that he had in receiving early funding from government and academic sources rather than from venture capital, which meant there was no pressure to create a financial event for his investors. As a result, he has been able to successfully steer his company through many business cycles while avoiding significant layoffs that are de rigeur among so many major companies. This has been a cultural differentiator, and is one of the reasons that SAS is regularly chosen among the best places to work in the United States.
(To listen to an unabridged audio version of this interview, please click this link. This is the fifth article in the IT Influencers series. Past interviews with Salman Khan, David Pogue, James Dyson, and Walt Mossberg can be accessed through this link. To read future articles in the series, please click the “Follow” link above.)
Peter High: You began SAS while you were an academic at North Carolina State. What was the genesis of Statistical Analysis System?
Jim Goodnight: When I was a sophomore, I took the only course in programming offered at NC State. The summer of my sophomore year, I had two jobs in programming. In the fall, I went to work for the department of statistics and it had a group of statisticians that were referred to as experiment stations. The experiment stations helped design and analyze all the agricultural experiments that went on on-campus.
NC State is a land grant university – there’s one in every state and the entire Southeast association of experiment stations sent one or two people each year to an annual meeting where they’d discuss computational methods and experiments that they’d been doing. NC State was sort of out in the front of developing analytical software and in the entire Southeast experiment station people decided just to use what was coming out of NC State at the time. Back in late 1966, early 1967, after the IBM 360 came out, everybody at all these universities was buying these IBM machines. They again looked to us to develop software for those machines.
Tony Barr started some of the first parts of SAS and then I joined him once it was stable enough to start writing additional procedures. We announced SAS in 1969 to the experiment station users – they were the group of university statisticians of the southern experiment station and they all liked what they saw and they started using it and we went on developing SAS. I was finishing up my Masters and working on a PhD at the time so I was working about 30 hours a week on the project. Back in 1972, I finished my PhD and we lost all of our funding from the NIH. The NIH, up until ’72, was providing funds for almost every computing facility at all the different universities around the country. Nixon decided he wanted to only spend money on universities that had hospitals, cancer research and things like that so you had to have a medical center from then on for NIH to provide funding. So we went back to our university statisticians group and said “how about supporting us?” They each chipped in $5000 a year to support us and they insisted that we also start licensing our software to other companies and other government agencies so we started doing that to become self-sufficient.
That went on until about 1976 at which time we had a user conference down in Florida. Our users actually put the conference together, but we went down and there were 350 users there. We were most impressed – they really liked some of the stuff we were working on, so when we finished with it and came back we decided to get out of the university. We were not able to grow anymore, there was clearly a lot of interest and we could support ourselves without needing the university. Of course the university, at the time, was not supposed to be a business anyway so they thought it was a good idea that we move off-campus as well – so we moved across the street and that’s how SAS was founded. It had a long development history at NC State before we left – I think we left with about 300,000 lines of code; it’s probably 10 million lines of code now.
Additional topics covered in the article include:
- SAS has a unique culture that has been lauded frequently, and considered a source of competitive advantage. How did that culture evolve over the years?
- You have rejected opportunities to be acquired. You’ve not elected not to go public either, whereas many of the technology companies that are peers of yours, in terms of size at least, have elected to do so. Can you talk a bit about why you’ve elected to stay private and the advantages of doing so?
- You echoed in your sentiment there that you can take a much longer term approach as a private company than a public company might otherwise do. You’re not a slave to quarterly earnings, you don’t have to worry too much about the natural ebbs and flows that the business cycles will exert on a company like this. This may not mean as dramatic a response as your peers among public companies might take. Is that fair?
- I also wonder how your responsibilities have evolved as the company has grown. What are you doing today that may be different from what you did years ago?
- As a former academic and as somebody who taps into the wonderful universities in this area and others you mentioned, could you comment a bit about the state of computer science in the academic realm. There are some who lament the fact that the areas that SAS, for instance, is strongest – analytics – isn’t necessarily a discipline that is taught very well at the university level.
- So given the number of universities you’ve mentioned that you’ve sponsored or others still who’ve probably elected to go your direction, are you generally optimistic about the quality of engineering or computer science degrees from here forward?
- If one has the pleasure of driving through this really beautiful campus that you have here, one of the first things, at least I noticed on doing so this morning, were the number of flags flying with the word analytics on them and clearly for some time that has been a very big priority for this organization needless to say. It’s now, over the past few years, become perhaps the hottest trend on the minds of CIOs at a minimum and to a greater extent, CMOs, CEOs in fact as well. I wonder as an organization and as an individual who has been thinking about this since long before this was such a hot topic crossing industries, what advantages SAS has relative to others who are newer to the game?
- As your business has evolved and as the users of SAS have changed from some of those early adopters among the pharmaceutical companies and financial services organizations and now to a much broader set of companies, how has the customer within those companies changed? Is the average customer now a Chief Information Officer or a Chief Marketing Officer or some sort of analytics role within the organization? Is there any way to sort of typify that and how that has evolved?
- Given the diversity of executives, just among the things you just mentioned, that could be a Chief Information Security Officer or it could be a Chief Risk Officer, you mentioned the Chief Marketing Officer as a user of data processing tool. Now as data has implications and there’s a recognition of the value of it across such a broad set of executives, does it make the sales process more complicated that there’s not a specific individual within the company that you can target or does it actually make it easier since there are so many people who are cognizant of the need to use data in a better way?
- You’ve also built a web of technology partners: Teradata, EMC Greenplum, Oracle, Accenture; as you think about developing an ecosystem of partners to help you do what you do, what’s the criteria that you use?
- You’ve always advocated investing a significant percentage of profit back into R&D. Where have you focused your R&D spend?