Ambassador of Analytics

In a wide-ranging interview, INFORMS President-Elect Anne Robinson discusses her passion for O.R., her view on volunteering and the “buzz” that’s igniting the membership base.

By Peter Horner

INFORMS President-Elect Anne Robinson

If not for a chance encounter with a mysterious member of INFORMS and an inadvertently amusing yet intriguing video titled “Operations Research + You = An Exciting Career,” Anne Robinson might have never heard of operations research let alone become president-elect of INFORMS, the largest O.R. organization in the world.

The defining moment occurred in 1993 when Robinson was a 19-year-old sophomore at Acadia University, a small liberal arts school located about 45 minutes outside of Halifax, Nova Scotia. Robinson and three of her fellow students in Acadia’s honors math class were invited to attend a math conference at Bentley College in Massachusetts. The purpose of the conference, titled “What’s a Mathematician Like You Doing in a Place like This,” was to expose promising math students to the different career paths in math.

In the spirit of “completeness” and after her school chums who wanted to be accountants and actuaries had signed up for the popular sessions, Robinson’s professor and then mathematics department chair, Paul Cabilio, suggested that she attend the session on O.R., the one no one else had signed up for.

“Sure,” Robinson said. “What’s O.R.?”

Robinson attends the session on careers in operation research, listens to a presumed member of INFORMS (she never did get his name) discuss how great companies such as Disney are doing all kinds of interesting things with O.R., watches the INFORMS video, goes back to Acadia all charged up to be an operations researcher and starts taking optimization classes. She goes on to earn a master’s degree in management science from the University of Waterloo, as well as a master’s degree in management science & engineering and eventually a Ph.D. in industrial engineering from Stanford.

The next thing you know, she’s on the career fast-track at Cisco, working her way up to “director of customer value chain information and data strategy,” until she leaves last year to accept a position as “director of supply chain strategy and analytics” at Verizon Wireless.

Robinson’s meteoric rise in the business analytics world is matched by her rapid rise up the ranks of INFORMS. In 2008, just six years after serving as president of the Stanford Student INFORMS Chapter, Robinson was elected to the INFORMS Board as VP of Marketing, Communications and Outreach, a post she held until her election as president-elect last year. When she takes over the reins as president on Jan. 1, 2013, Robinson, 38, will be the youngest person in history to serve as president of INFORMS.

During the INFORMS Annual Meeting in Phoenix in October, we sat down with Robinson to discuss the prominent role she’s played helping INFORMS capture the analytics movement, her take on several key INFORMS programs and initiatives and her vision of the Institute’s future. Before the interview began, however, Robinson had a question of her own: If you are reading this interview and you were the person who gave that presentation at Bentley College in 1993, please let her know. She really wants to thank you.

You’ve been a highly visible and vocal champion of “analytics” since you first joined the INFORMS Board five years ago. What convinced you that INFORMS needed to jump on this particular bandwagon?

At the time, [INFORMS VP of Practice Activities] Jack Levis and I and other practitioners were seeing and hearing this big buzz around the word “analytics” externally, but there was little mention of it within the INFORMS community. I began identifying myself and my work using terms such as “advanced analytics” and “business analytics” – and that’s how people at Cisco started identifying my team – yet within the INFORMS community, we weren’t embracing it. I saw a tremendous opportunity for INFORMS to get involved.

How important is it for INFORMS to capture and capitalize on the analytics wave?

It goes back to being able to articulate what the field of operation research does. I started in math and then learned about operations research, but it was a challenge. When I first heard of it, to me O.R. meant “operating room.” As I got into the field, the words “operations research” still didn’t resonate with many people outside of the field.

Five years ago, during my first month on the INFORMS Board, I recall talking with you and others on a conference call about starting an executive magazine as an outreach activity for the organization and what we should name it. Click: “Analytics.” It took us about 10 seconds to settle on the name, and that was the start of it. Capturing that title at the time that we did was probably a lot more foresight than either of us realized.

Tom Davenport’s 2007 book, “Competing on Analytics,” is widely credited for building the analytics buzz.

Exactly. We had this person out there, Tom Davenport, who had written a book that was getting so much attention, creating so much buzz in the business community, and I’m thinking, “This is what I grew up in; this is what I do.” Why aren’t we – meaning INFORMS – bringing operations research and analytics together?”

I then started going through the process of being an advocate for analytics to the Board, and putting it in the context of what I as a practitioner was hearing and what my peer practitioners were also hearing and how they were self-identifying. It was a natural transition that took some convincing, but what business people were looking for as far as analytics was concerned was what INFORMS brought to the table.

Getting out in front of the “analytics” wave is now an important part of the Board’s strategic plan for INFORMS.

I’ve been around long enough to see the challenge with the name “operations research.” I’ve seen people struggling to explain operations research and struggling with their professional identity because of the name. Here was an opportunity to change that. Now those same people – even those who said analytics didn’t necessarily resonate with them – are using terms like descriptive, predictive and prescriptive analytics to describe themselves and what they do. The evolution has happened right in front of our eyes these past few years.

Once convinced to go forward, the Board made it clear that it didn’t want INFORMS to be just a player in the analytics space; it wanted INFORMS to be a leader, particularly in what it called “high-end analytics.” What makes INFORMS suited to play such a prominent role in the analytics movement?

Analytics is so much of what we do that it only makes sense for INFORMS to lead and embrace it. On the flip side, one of the things I’m noticing and one of the reasons I’m so keen on this advocacy- or ambassador-type of perspective is this: If you go to Predictive Analytics World or a similar event today, often times when they speak about analytics they end it at predictive analytics. They go as far as understanding how to forecast, but they are not getting to the prescriptive space, which is really our sweet spot.

I understand you met Tom Davenport early in your working career and your story appeared in one of his books.

As I was starting my journey as a manager of the analytics function at Cisco, I went to one of SAS’ “Competing on Analytics” executive events where Tom was speaking. I met him afterward and we spoke for a few minutes, and somehow my personal journey into analytics struck a chord with him as he reached out to me several months later.

Let me back up a little bit. When I was starting the new forecasting function at Cisco, I had the good fortune of meeting with Scott Ellis, who at the time was the INFORMS Roundtable member representing Hewlett-Packard. He said to me, “If I could do it all again, I would do analytical model development as rapid prototypes and get stakeholders involved at every step of the way.” I was just starting out in my career and here’s someone who had been very successful, so I decided to take his advice, and follow what he had said he would do. This is one of the great benefits I found as a practitioner in INFORMS. I had access to people who had actually been there. I thought, “This is my leadership group, this is my peer group.”

I leveraged that wisdom from Scott and did rapid prototyping at Cisco where we first tested forecasts for a small number of products and using basic models. We got stakeholders involved, iterated on the models and expanded the product, and went on this journey together. Within a very short period of time, we developed a full advanced statistical forecasting capability appropriate for Cisco’s business.

That story somehow got back to Tom, who sent me an e-mail. He told me he was doing new research, resulting in a piece title “How Organizations Make Better Decisions,” and he wanted to talk to me about it. We met, we talked again, and my team’s journey became one of the case studies that he wrote about. He and I have remained in touch over the years. He’ll reach out to me on occasion and he often mentions INFORMS when he’s talking about the landscape of analytical organizations.

Early on during its own analytics journey, INFORMS hired the consulting company Capgemini to assess the analytics marketplace and to outline opportunities for INFORMS to exploit it.

Capgemini researched the situation and came back and said, yes, there’s a market out there and, yes, there’s an opportunity for INFORMS to meet that market need. Their report also indicated a sense of urgency because other groups were also looking to move into this space. One piece of research we found particularly interesting was that operations research was seen as a toolbox for solving a problem, whereas analytics was seen as a business capability. Capgemini also came back with a full breadth of opportunities, including certification, continued education, more executive offerings – all initiatives that we are now making real progress on.

Of those initiatives, the Certified Analytics Professional (CAP) program is the furthest along, with the first set of exams scheduled in conjunction with the INFORMS Conference on Business Analytics & Operations Research in April 2013. Why is CAP so crucial to INFORMS and the profession?

With this groundswell around analytics, there is a shortage of analytic talent out there, and that talent shortage is increasing. I’ve personally experienced this shortage as I’ve been trying to hire people.

Currently, there is no existing certification on the end-to-end analytics process. We see this as a way for people to promote and differentiate themselves in the job market. For INFORMS, it’s a way of branding and associating analytics with INFORMS. It’s also a means for INFORMS to bring in new members and potential revenue in a way that it has not done before.  

The benefit for analysts who go through the CAP program and pass the test is they’ll have a Good Housekeeping-like seal of approval from INFORMS.

Exactly. And the reality is, externally, people are beginning to recognize that INFORMS is a source of expertise for analytics.

As you alluded to earlier, defining “operations research” has been a sore spot for more than 60 years. Within a few months, INFORMS came up with an “80 percent” solution in defining “analytics.”

That was what did it, the 80 percent solution. It was important to have a definition because people kept asking, What do you mean when you say “analytics”? I realized that the first thing we needed to do was own a definition. Initially, we didn’t attempt a definition; in fact, we originally said, “Just define it how you like, just think of descriptive, predictive and prescriptive.” That’s what resonated, first. But once that was out there and people were turning to INFORMS more and more and looking at us as a source for analytics, members were asking about analytics, and we realized we needed our own definition.

We also recognized the personality of the O.R. community and the fact that you’re never going to get a 100-percent satisfaction definition. We like optimality! So we went through several iterations and came up with a definition that 80 percent of those polled agreed on.

So what’s the definition?

The scientific process of transforming data into insight for making better decisions.

That sounds a lot like operations research.

The “scientific process” piece was meant to infer the depth of knowledge around analytics. This wasn’t about defining operations research. We needed an internal definition of analytics that was simple – everyone advocated for something simple – and something that resonates. It was about “transforming data into insight.” That was the big piece, and the word “transforming” embodies what we do.

The definition seems pretty broad and covers the full spectrum of analytics. Was that by design? Isn’t INFORMS focused on the so-called high-end: predictive and prescriptive analytics?

The definition is broad. Do we go after every part of it? Maybe, maybe not. The idea was first to get a definition out there that resonates. When you look at that full landscape, if you look at DWI for example, the Data Warehouse Institute, when they talk about analytics they are really focusing on data bases, master data, data schema, what that looks like, how to establish the data sets, getting into business intelligence capability. There are a lot of players in that space right now, and I don’t know that it makes sense for us to get into it as well.

The number of groups in the higher predictive and prescriptive spaces are much lower, so that was a reason for saying, as much as we have a huge appreciation for data management and the importance of bridging IT and analytics, I don’t know that we need to bring that full community under the INFORMS umbrella.

I can see the advantage of INFORMS having a big tent and welcoming analysts all along the maturity chain. At the same time, INFORMS has roots in operations research that go back more than 60 years that you have to respect and protect.

O.R. is still at the core of INFORMS and always will be. We have the top analytics in the space and the top academics in the space, and I never want to lose that. To help people and companies understand and benchmark where they are in terms of analytics, we are creating an analytics maturity model.

Can you talk a little more about the maturity model?

It’s still in the development phase, but we’re looking at giving an individual or a company the ability to self-assess where they are in terms of their use and deployment of analytics, perhaps through a multi-step questionnaire, and then helping them move up the ladder. At the same time, I don’t want us to miss out on the fact that some of these companies are going to be ready now for that O.R. capability, prescriptive analytics, and let’s help take them all the way to the top.

I think some of the language out there has suggested that analytics is going to potentially open our community to a less rigorous capabilities than we are used to. To the contrary, I think it will make people aware of our toolbox and skills and unique way of looking at complex problems, and we are going to have a group within this potential new member base who are going to be hungry to learn and acquire those tools and skills. I think it will strengthen the O.R. position rather than diminish it. The future in this topic area is limitless.

Continuing education is another key INFORMS initiative.

Steve Powell and Rina Schneur have been leading the committee on this for the past eight or 10 months. Employers are looking for a certain skill set. INFORMS is 70 percent academic, yet there remains a big gap between academia and practice. We’re trying to fill that gap with continuing education.

Regarding all of these initiatives, INFORMS has conducted numerous surveys, hired consultants and done due diligence, but there’s risk to go with the potential reward.

I’ve encouraged every member I’ve spoken with, whatever your perspective – love it, hate it, whichever – to give us your feedback. It has to be right for our community, and if the community wasn’t supportive, then it wouldn’t be the right direction to go. The support for the analytics-oriented initiatives has been amazing.

However, on the flip side of that, one of the beauties of operations research is that the results are often not intuitive. One of the things I say to executives when I’m talking about analytics, particularly prescriptive analytics, is that it’s somewhat high-risk, high-reward. You have to be brave enough to take the risks on the results because it’s likely to be counter-intuitive to what you are doing today. I think our community needs to use that same philosophy with analytics – let’s be brave enough to take the leap!

What do you tell the skeptics within INFORMS who might question the Institute’s considerable investment of resources in the analytics movement?

We are always committed to keeping operations research at the core. We never want to disrupt that golden nugget that our society has, that depth of expertise. But the reality is, the society was not doing well in terms of membership growth, to a point where you could almost predict when that last member was going to turn off the light. We tried different things in the past to push out operations research, including videos, press releases and public relations, but the reality is they didn’t resonate. This is an opportunity for us to embrace a wave that we fit into, one that’s re-energizing the base. There’s a buzz among INFORMS that wasn’t there the last few years.

I know some people are a little skeptical, but we have no plans to take anything away from our fundamental O.R. core. That is our history, that is our identity, and I want that to last. I’m at a different point in my career than some of the past presidents have been. I still expect to be doing this 20 years from now, and I want INFORMS to be there with me. I want to see us stay and accelerate and grow. This is one way to do that.

It should be noted, as outgoing INFORMS President Terry Harrison pointed out during the Membership Meeting at this conference, that INFORMS is on solid financial footing.

Absolutely. We had a surplus last year, which has allowed us to pursue some of these strategic initiatives. INFORMS has been very healthy financially, which is fantastic. We positioned some strategic money to devote to these initiatives. The reality is we had a surplus last year that funded everything we wanted to do. This year we’re pacing toward a surplus again – membership is up and attendance at the conferences is as high as ever.

You mentioned meetings. I understand total attendance at this meeting is right around 4.500, which puts it close to the all-time high. Meanwhile, attendance at the analytics-oriented spring conference jumped by more than 30 percent this year.

The meetings are doing very well. Not only the two flagship events you mentioned but also the topical meetings. The 2011 healthcare conference did so well we are doing that again in 2013. In 2014, we’ll hold a meeting on big data. In addition, we’re contemplating offering another kind of event or meeting to coincide with the certification program, and our chapter, regional and subdivision meetings are doing quite well.

After years of slow decline, INFORMS membership jumped significantly last year, but most of those new members are either students or first-time conference attendees, and both groups received huge discounts on their initial membership fees, if they had to pay at all. How do you convert them into delighted, long-term, dues-paying members?

There is a churn, particularly with new members, but I think the regular member retention rate is around 80 percent. One of the things we learned from a member survey is that members who join subdivisions and communities tend to have a greater “stickiness” because they feel more engaged with the Society. If you’re involved in the transportation industry, for example, and you join the TSL (Transportation & Logistics Society), you’re more likely to renew your membership because you’ve found your community and your home within INFORMS. One of things we’re trying this year, for new members in 2013, they’ll also be given membership into the subdivision or community of their choice as part of their regular membership.

You were a member of the Roundtable, which is essentially a community within INFORMS. What impact did that experience have on your career?

I was fortunate as a student volunteer to be invited to sit in on a Roundtable meeting where I was exposed to this group of people who had all essentially gone along the career path that I aspired to. It was an amazing experience.

Later, as a junior practitioner, I convinced my company to join the Roundtable, with me as the representative member. From that point, at every step of my career, members of the Roundtable helped me learn and understand how to be successful, based on their experiences. In most companies, as an advanced analytics practitioner, you are often one of two or three, if you’re lucky; in smaller companies you might be the only person. That’s especially true if you’re an analytics manger; you rarely have a peer group. I found myself learn as I went, and really credit the Roundtable for enabling me along the way.

I remember having a long conversation with [Roundtable member] Russ Labe on soft skills and how to train an analytics team. I can remember having conversations with [Roundtable member] Glenn Wegryn about how Procter & Gamble approached analytics and some of the ways that he had accomplished success and what I might do in terms of crafting and growing an analytics capability at Cisco. It was really this group that taught me how to lead and how to manage and mentor. I’m extremely grateful for having had those experiences through the Roundtable.

When you talk about the value proposition of INFORMS membership and you talk about networking, I don’t think people quite realize the power, the knowledge and the guidance that can come from other members such as those I found through the Roundtable.

Volunteering is another way to get members engaged. Can you talk a little about the role of volunteers in INFORMS and your experience as a volunteer?

I’m very grateful to everybody who has volunteered their time to INFORMS. Being a non-profit, member-based society, we really represent a collection of volunteer energy. We have a fantastic staff, but INFORMS would not exist without its volunteers – whether they are working on awards, crafting new ideas, fostering a student chapter or doing whatever you happen to get involved with.

My reason for initially volunteering was to gain exposure to these great people in INFORMS that I never would have met had I not volunteered. Early on, as a student volunteer, I remember meeting Art Geoffrion. I had read his papers in class and recognized him as a thought leader in our field, and here I was sitting next to the man himself. It was a great experience.

I would encourage everyone in INFORMS to reach out and volunteer. It’s a great way to meet people. If you have a new idea, do not be shy to suggest it. Get involved. There are so many ways INFORMS can help align your volunteer work with your professional, personal or academic interests and passions.

When you’re not directing supply chain strategy at Verizon Wireless or working as a volunteer for INFORMS, what do you like to do for fun, assuming you have any time left for that sort of thing?

I have a passion for music. I grew up a classically trained pianist, and I play a lot of instruments. I’m a big fan of jazz. I sold my piano when I moved from California to the East Coast, but I’m looking to buy another for my new place.

I started running last year and I’ve run several half-marathons. I don’t know that I enjoy the running part, but I love the excitement at the start of a race and then the satisfaction at the end. I love discovering and trying new things. While living in California, I learned how to surf and, thanks to my Ph.D. advisor Bob Carlson, how to properly taste wine. I enjoy biking. I guess as a child you try to be an expert at a few things, and as you get older you just want to be able to try everything.

What do you hope to accomplish during your year as president and where do you see INFORMS five years from now?

At the end of next year, I would like INFORMS and analytics to be synonymous so that when business decision-makers think of analytics, they think of INFORMS. All these initiatives around analytics we’ve discussed, including members becoming INFORMS ambassadors for analytics, I want to see expand and flourish.

Five years out I want to see an institute with a very strong analytics backbone that is really seen as driving business decision-making on a global scale. I can see our practioner base stronger than it is today. I hope all these initiatives that we talked about become a fundamental part of the core offerings from INFORMS, with many certified analytics professionals a part of our community.

You can feel the energy in the air as this conference is buzzing with analytics, and I want to see that groundswell continue now, next year and well into the future.

Peter Horner is the editor of OR/MS Today and Analytics magazines.