Building and Operating a Modern Analytics Center of Excellence
This post is contributed by Kathleen Maley, VP, Analytics Executive and Member of the IIA Expert NetworkSo you’ve decided to get serious about making your business more competitive. You’re finally ready to squeeze more out of your investment in data analytics. You’re ready to elevate your analytics team from order takers who merely serve up requested data like waiters in a restaurant to thought leaders who sometimes challenge preconceived notions and help business leaders find better information that helps them make better plans.Or at least you’re thinking about it.If you aren’t at that stage yet, then stop reading now. The rest of this article won’t interest you because you’re not ready for it yet, and it won’t be a good use of your time. But if your business has matured to the point where you can see the limits of on-demand data retrieval, if you see the need to make data analysts partners in setting strategy -- but you’re just not sure how to do it -- read on. You will find the next several paragraphs thought-provoking and possibly transformative. I will lay out for you the six best practices for evolving your analytics Center of Excellence (CoE). By doing so, I hope you’ll be able to see how your organization might use these practices to operate more efficiently, gain a competitive edge and make more money.I know how hard it can be for analytics leaders and business leaders to work together effectively. I’ve been on both sides of the equation. I started my career as a member of the Analytics Center of Excellence at Bank of America. I later led the analytics CoE for a regional bank in the Midwest. I’ve also been on the business side, running my own profit-and-loss center and leveraging talent within an analytics CoE. As a result, I’ve developed a unique perspective on how to most effectively establish and work with an analytics CoE. I’ve learned that creating value from analytics is a two-sided opportunity and each group -- the analytics group and the business line -- has a very specific role to play.I started out just like most analysts -- formally trained in statistics, a bunch of useful algorithms and various programming languages -- but no one ever taught me how to convert that knowledge into meaningful return on investment (ROI). That’s because most people don’t know how to do it. And guess what? No one ever explained to my partners how to work with me, what to expect from me, nor what standards to have for my work product -- because nobody knew any of those things. I had to figure it out myself.Most analysts and business partners still struggle to talk and work with each other in a way that effectively leverages the expertise of the analytics group, so below I’ll cover the six main best practices that can help. None is a silver bullet and each needs to fit within your own company culture, but these are the major areas to consider as you evolve your analytics capabilities.
Agree on purpose and approach of the CoE
First and foremost, evolving your CoE is about developing strategic partnerships. If both parties don’t explicitly buy into the objective and agree on a set of operating principles, it’s not going to work.Start by bringing people to the table, having an open and honest conversation and then linking arms to head into the future together. This is a really important step and it’s also sometimes a difficult step because the analytics leader likely has to ask for trust that she might not yet have fully earned. Asking someone to step into the unknown requires that person to take some risk, so it’s critical to be empathetic, maintain open communication, and productively embrace feedback along the way.
Organize your team
Generally speaking, I favor an organizational model that is centralized and fully dedicated. As the head of an analytics CoE, I’m responsible for making sure that I hire the right mix of specialized skills, that I can effectively manage the quality of my team’s work product, that there is career path, employee engagement, and all of the other things that go along with talent management. But they don’t actually do any work for me.All of their work, almost without exception, comes from the business partners they support. My primary responsibility is to make sure my partners have access to the best talent and then set the stage for effective engagement. The “fully dedicated” part is about evolving beyond pure fungibility of resources toward resources that are consistently supporting the same business. This facilitates relationship development, and enables analysts to develop depth of knowledge and history with a specific business line.
“Who do I go to when?”
I recommend a single, primary point of contact (POC) whose job is to build a relationship with a specific partner or set of partners. That POC must become deeply steeped in the partner’s business and strategies, manage the portfolio of analytical work being done on behalf of the partner, and make it easy for the partner to access the entire analytics ecosystem.That last part might be trickier than it seems because the analytics ecosystem is growing more complex. There are automated insights, business analytics and predictive modeling. You also have functional business questions to contend with. For example, if I am a product executive and I have a question about digital acquisitions, do I go to the analyst who supports product? Or the analyst who supports digital?
Answer: It depends… And because of that, it will always be easier for the primary POC to navigate the analytics team in order to leverage the analytics ecosystem holistically.
How it Works
There are several specific operating rhythms I recommend, and you can think of them as different gears in a system. They include a business strategy session, a portfolio-management routine, an internal portfolio review and periodic leadership updates.The business strategy session is used to set top-level priorities once or twice a year. I have typically done this by running half-day sessions that include a business-led deep dive on their strategy and current priorities, a POC-led review of the existing portfolio of analytical requests, and an alignment of the partner’s top priorities to the existing portfolio. During this part of the session, new requests are added to the list and existing requests are reprioritized or dropped from the list entirely. Note, however, that it is the partner who decides priorities. The analyst will guide and make recommendations, but at the end of the day it is up to the business leader to decide between short-term and long-term needs.The portfolio-management routine is run by the POC on a bi-weekly or possibly monthly basis. New requests are prioritized, conflicting priorities are sorted out, and roadblocks are knocked down. This routine is not a report-out. This routine is for the POC who uses it to ask clarifying questions, raise risks, and facilitate decisions with his partners so that the work continues in a meaningful way.The weekly (or bi-weekly) internal portfolio review has two main functions. It allows the analytics team to break down silos for themselves and for their partners by talking to each other about the questions they are getting and the work they are doing. This is one forum in which I start to challenge my team to develop new muscle memory by asking them WHY the request was made -- what problem or opportunity is the partner trying to solve. As a CoE evolves from “getting the data” to supporting effective business decision-making, everyone involved -- analysts and business partners -- has to shift their mindset, and that takes deliberate intention. I use the internal portfolio review as a safe space to ask the analysts “Why?” In the early days, most people can’t answer that question. So without judgment, I ask them to go back to the partner, have a conversation about what the partner is trying to solve, and then come back to next week’s review with an understanding of the purpose of the request. Almost without exception, understanding the broader purpose is Step 1 of being a “thought partner”.Senior leadership updates are less important for the day-to-day work but of significant importance in maintaining the health of the CoE as a whole. I’ve done this in a couple of ways -- monthly summary by email or PDF, standing agenda item on a leadership routine, or some combination of the two. Either way, I approach it as a shared update that my partner and I both put our names on. I share what we found in the data, and my partner shares what he’s going to do with it and what it’s worth. By sharing, we demonstrate the partnership and create accountability for both of us, because the probability of action increases significantly once we commit on paper.
Deciding what is important
Prioritization happens at several levels: within a business unit, across business units, and enterprise-wide. Within a business unit is the easiest and happens most frequently via the portfolio-management routine described above.Prioritization across business units usually occurs when a leader has a significant opportunity but no available baseline analytical support. In that case, we might jointly approach another leader and have an ROI-based discussion about “borrowing” a resource for some period of time.Enterprise-level prioritization is the least frequent, and is aided significantly by the semi-annual business strategy sessions. That is the time to look at all of the business opportunities at play across the enterprise, the potential ROI of each, and jointly decide where to align resources. This usually involves operating at a higher level of leadership.
Paying for it
I favor a hybrid funding model, which involves providing each business unit with some baseline level of analytical support that is centrally funded as a shared service CoE. If a business leader needs additional support that, for whatever reason, can’t be provided from the baseline pool, that leader can choose to fund an “untouchable” resource from his own budget. It’s very common when I’ve taken over a team to hear, “I can’t get what I need from the analytics team. They are too busy and my work falls by the wayside, so I have to hire my own people.”To that, I’d say, go back and try putting some of these routines into place. Second, I’d say, “You might be right, so fund a resource that is 100% dedicated to you, but let me use my expertise to help you hire the right person and provide that hire the right talent management and employee engagement routines.”
I hope you can see how all of these routines work together to create a higher-level of operational partnership. One, we collectively know the purpose of a centralized CoE. Two, everyone knows his or her role in doing the work, managing the work and prioritizing the use of this valuable analytical talent. And last, we know where higher-ordered tradeoffs need to be made AND we make sure that those decision makers have the information and connectivity to the work -- and to each other -- to do so.
About Kathleen Maley
An experienced analytics executive and P&L owner, Kathleen Maley specializes in making data and analytics work for organizations that are challenged to bridge the divide between data and business value. By combining strategic vision, technical expertise and highly developed emotional intelligence, she effectively guides the evolution of data analysts and business leaders into thought partners and data-enabled decision makers. Kathleen is a published writer and frequent speaker and holds degrees in Mathematics (BA) and Applied Statistics (MA).