BI attracts the naturally curious
The explosion of the software industry had its roots in the shift from mainframe-style computing to client-server architectures. The shift from mainframes and mini computers to Unix in the late 80s, was a shift in the industry that changed everything, although it was hard to see at the time.
So says founder of Third Nature, Mark Madsen. "The same thing is happening today, only now I'm looking at it from the other side. The mainframe batch systems of then are the client-server database systems of today. The arguments I had with old mainframers then are the arguments of a new generation today with the IT establishment.
"The new people say that cloud and stream-based architectures will replace the client-server model. They're right of course, but most of us don't like it any more than COBOL programmers liked Unix and databases in 1992. The wheel turns and eventually comes back to where it started. The only thing that changes is the people."
Madsen, who will be presenting at the ITWeb Business Intelligence Summit 2017, to be held on 14 and 15 March at The Forum in Bryanston, says he got into the business intelligence (BI) field by accident. "I had been working for a Nobel laureate in economics. What I learned on the job was applied to management decision-making. When that didn't work out - people don't like being told they made an irrational decision and they find something else for you to do - I discovered that the technology side of the work was interesting and challenging in a different way. It didn't hurt that it paid about three times a university salary."
A move to IT
On his move into the IT field, he says he was working full-time as a research programmer in the cognitive psychology department at a university. "After I finished my dissertation I realised I'd have to spend another three years in graduate classes while working full-time in order to finish my PhD (in was in autonomous mobile robotics). I decided to accept a job offer at a manufacturing company instead."
Before the university job he did a lot of C and assembly programming. "After, it turned out that applying behavioural economics to decision-making wasn't that practical or welcome back in 1992, but this thing called decision support, later renamed business intelligence, was starting. It was a small shift from behavioural economics to this line of work. I did this first at the manufacturer, then at a parallel systems company, then as a consultant. I moved to work at Internet startups for a few years, ending up as a CIO and then a CTO before going back to the BI field as a director of analytics at a multinational company."
From there Madsen chose the independent route, and left the BI market for the second time. "I was working on real-time telemetry and large-scale data analysis for a few years. That made me unlearn a lot of what I thought I knew about data and analytics and BI. Today, the big data and analytics market and the data warehousing market are merging into something new. I've been adapting the new market architectures to the old BI and data warehousing problems, and vice versa."
Good and bad
In terms of what he loves about his work, Madsen says he enjoys the variety. "Each project is in a different market, industry or country. Technical problems vary. Some people like stability. They like to make sure things run smoothly. It's a vital operations role that takes an aptitude I don't have. I found that I didn't make a great CIO because I get bored if I do the same thing too many times. I like novelty. With novelty comes risk, and corporations generally do not like to take risks."
On the other hand, he says the frustrating part of his work has mainly to do with the tension between people's desire for stability and their desire for improvement, which implies change. "I'm often hired to do something new, or to clean up a mess. Sometimes those two things go hand in hand. But there is always resistance. It's natural to fight change, but in the end that is self-defeating. The hardest part of the job is convincing people that they need to change their assumptions, what they believe is right. Unlearning what you know is a hard thing."
Were there any defining moments in his career? "It was probably holding out for a job I might not get but was really interested in, when I had a good, safe offer in hand. I made the decision because of the people I'd be working with. I got that job. I learned that who you work with is far more important than the type of company, or industry or location or even the specifics of the work."
However, he says that's the answer assuming he believes in single defining points. "I don't, really. Life is full of random events, little decisions that change your direction in big ways and carefully considered decisions that don't."
Madsen says, for example, the most useful knowledge he has came from the liberal arts, not from his science background. "What I learned of sociology and ethnography was far more useful in BI than technical skills. Film school and art history had more impact on successful visual design and communication than knowing how to code."
One of the things I picked up in a past role was that interdisciplinary knowledge is as valuable as deep domain knowledge. "Other domains have solved the same problems in different ways. Who would have thought that Sumerians struggled with the same information management problems we struggle with today, or that we could learn something from them? So I believe in serendipity more than a clear defining moment."
Fun and interesting
Would he have done anything differently? "Big decisions always seem smaller in retrospect. The risks that seem so large turn out not to be as large as you think. I would not have hesitated to change jobs or take projects where I didn't feel I knew enough to succeed. It turns out that being uncomfortable is when you learn the most, and often do the most interesting work."
If he had to describe his career in one sentence, Madsen says: "It's a cocktail party that lasted all weekend because the people turned out to be much more fun than expected."
The other thing Madsen learned is that there are parts of the IT industry that are fun and parts that are not. "Working with data turns out to be a fun part of the industry. Maybe it's that curiosity is a useful attribute and so the field attracts the naturally curious. Or maybe it's that data is unpredictable, and that attracts people who are comfortable with uncertainty. Those two attributes could lead to a more relaxed culture than you find in other areas of our business."