The Analytics CoE: Don't Get Smashed on the Rocks

If youíre thinking about creating an analytics Center of Excellence (CoE), Iíd advise you to slow down and exercise extreme caution, because it can be a sirenís song -- many are the bodies of analysts smashed upon the rocks of the CoE.

On its face, the CoE has great appeal. After all, it has "excellence" in its name, and who wouldnít want to be a part of that? And anyone who has spent time in analytics knows the issues that organizations struggle with as they mature in their analytic capabilities. A lack of process, no single version of the truth, minimal transparency, and an overlap of spend on analytic resources are common problems that the CoE is designed to address. But the CoE brings its own set of problems, while also exposing your company to considerable risk that comes from rearranging your decision-making infrastructure.

You must recognize that when you consolidate your companyís analytic resources, centralize your data, and create standard processes and best-practices, you are tampering with the language and culture your company uses to make decisions. This happens any time you change the way information flows within an organization, and the more radical the change, the greater this impact. And if thereís one thing that organizations donít like, itís change. They abhor it, they fight it, and they will work to defeat it. Even if they recognize that itís good for them.

So you better be sure you know what problem youíre trying to solve with the CoE. You better have executive leadership to drive this change deep within your company. Your company better have the maturity and experience to navigate this complex and difficult transition. And you better design your CoE to be obsolete in a relatively short period of time.

When you know what problem you are trying to solve, you give clarity to yourself and your organization. So you must answer the question: What will we do differently as a result of this change? Do not dare ask your company to make this change if you canít answer this question with precise clarity. Not only does answering this question give you focus, it also gives you the rallying cry for your organization to buy in to this cultural change.

But no matter how clearly and precisely you can answer this question, donít ask your organization to make this change if you donít have an executive sponsor eager to lead this effort. Donít fool yourself that if you build a CoE then the rest of the company will fall in line. No matter how well youíve designed your CoE, no matter how much sense it makes for your company, no matter how much money youíll save, or how much youíll improve your decisions, you wonít get anywhere without a very senior executive who is ready, willing, and able to expend political capital to inspire, cajole, nudge, poke, drag, and whatever else it takes to make your organization operate differently.

Ideally your executive sponsor is your CEO, or perhaps another C-level executive. Depending on your organization, you may be successful with a CoE sponsor who reports to a C-level executive. If your sponsor is any lower in the organization than that then its likely he or she wonít have the authority to effect the kind of change your CoE is intended to create.

But even clarity of purpose and an outstanding executive sponsor isn't enough to guarantee success. For that you also must have an honest assessment of whether your company has the maturity in its analytic capability to make use of a CoE. You need the basic building blocks in place, including some sort of data warehouse with multiple groups reasonably adept at analytics. In other words, if youíre living mostly in Excel and Access, you arenít ready for a CoE. Having established analytic strength goes a long way in helping overcome the institutional inertia you are going to face. These silos of analytic expertise provide the individual success stories that will help sell your CoE to the rest of the organization. If you donít have these, then donít waste your time on a CoE -- youíre not ready. Instead, focus on building analytic strength in one or two silos.

If you are able to cross over all of these hurdles, then as you design your CoE, you should do so with a limited lifespan in mind. Thatís right -- just like the replicants in Blade Runner, your CoE should shut down after a period of time (two or perhaps three years at most). Thatís because the creation of a CoE in some sense disconnects decision making from decision makers.

If your company is going to have a true analytically-driven culture, the systematic use of data and analytics to drive decision making must in the long term become native to your company. Thus your CoE is merely a bridge between your current siloed state and your real goal of an analytically-driven culture. And so, while it may make sense to centralize your analytics in an incubator in which your analytic culture can be cultivated and grown while it matures, after a somewhat limited period of time this culture must become part and parcel of your companyís overall culture. Setting an expiration date for your CoE is a signal to the company of your long-term goal.

Lastly, should you reach the point where a CoE makes sense for your company, I beg you: Please donít call it a Center of Excellence. Itís an amorphous term that can be considered arrogant and insulting. It implies that prior to the CoE, your companyís analysts were something other than excellent. And it presumes a certain level of performance that has yet to be proved. Youíre starting on a journey with many headwinds, as you are sure to meet with much resistance from your companyís culture. For goodness sake, donít make it any more difficult! Better to give yourself a name that provides more clarity about what your value is, and isnít the equivalent of giving yourself a "gold star."

Have you built or attempted to build a CoE? Do my ideas resonate with you or are you more aligned with Sandra Gittlen's Counterpoint stance? Share your thoughts on the message board below.

Point / Counterpoint, Analytics Director, Hospitality Industry

Counterpoint: Train for Data Visualization Skills

Chances are you've already got good data visualization experts on staff, even if you don't know it yet.

Customer-Centric Banking Analytics Scares Me

Banks have tons of customer data at their disposal; unfortunately not all will use it scrupulously.

Re: Addressing the business need
  • 8/13/2012 5:47:39 AM

David I like your stance on CoE , and I agree that it should be used more as a transitional resource to give direction and cohesion to what might be a confused structure, but once these goals are achieved it is probably best not to adhere to CoE based principles too tightly.

Though there are good arguments both for and against the creation and use of CoE.  I tend to lean toward against or at least use as little as possible.

Re: Addressing the business need
  • 8/10/2012 11:01:34 AM

Rob, thanks for this insight. You clearly are passionate about the role of the analyst, and I hope everybody does indeed strive to find "an environment where your creativity and passion can thrive." Fortunately, I suppose, if the analytics talent shortage hits as hard as some predict, analysts may have the flexibility of changing jobs until the find that perfect fit!

Re: Addressing the business need
  • 8/10/2012 10:40:07 AM


I think that using analytics to further a chemical engineering competitive advantage would be great.  But, to David's point they'd need to be able to articulate with perfect clarity how analytics would do that and make sure it's a good fit for their organization and culture.

But I think merging the logistics analytics team into an overall CoE to support objectives other than logistics would not make sense.  It would be a bad fit from an alignment of goals and cultural perspective.  Strategic leadership wouldn't see the work of the logistics team as a priority and would neglect them.  In addition to being minimized and neglected by leadership the logistics team would probably also feel out classed by working shoulder to shoulder with a bunch of PhD Chemical Engineers.  At best this would lead to lower moral and less effort and enthusiasm by the logistics team and at worst could result in mass resignations and lose of an important logistics analytical capability.

Passion, creativity, and enthusiasm make a superior analytics team and leadership should make effort to enhance and maintain it.  You can buy a person's hands but you can't buy his/her heart.  Leadership has to earn that passion and creativity and there's no surer way to lose that than by under valuing and neglecting a team's contribution.  One of the analytics teams I worked for a year or so ago was hitting on all eight cylinders until a new leader was brought on board who didn't get it.  Now, in the short time frame of a little over a year 90% of that team has found another job and the team's metrics and moral have gone through the floor.  It isn't hard for a strong analyst to find another job (even one paying more money).  The harder thing is to find an environment where your creativity and passion can thrive.


Re: Addressing the business need
  • 8/9/2012 2:04:11 PM

Rob, thanks for joining in this discussion. What you say makes good sense, but consider this relative to your sample chemical company. ,Isn't it likely -- or shouldn't it be -- that its using analytics to make smarter decisions about which new chemicals it ought to focus its development efforts on, to predict customer demand for a particular chemical type, etc.? If that's the case, wouldn't those efforts, since they determine competitive advantage, be best served by a CoE? And if such exists, doesn't it make sense to bring in functional analytics areas, like freight, into that group for shared knowledge, collaboration, etc.? Just thinking aloud here ....



Addressing the business need
  • 8/9/2012 12:23:26 PM

Interesting discussion ...

In my view, whether or not to create a dedicated CoE (by whatever name) depends on alignment to core and compelling strategic goals.  A CoE, reporting to a C-level executive, would be appropriate if the company's competitive advantages depend on analytics.  For example, Netflix and IBM use analytics as a competitive advantage and it's hard to imagine those companies being successful without strong, centralized analytics functions and expertise.  I can also imagine other companies such as UPS benefiting from an analytics CoE with a specific focus such as optimizing their global supply chain. 

But without alignment to clear competitive advantage I think a CoE at the C-Level could very easily flounder due to organization politics and a culture of 'so what.'  Where analytics doesn't support competitive advantage the distributed model is more likely to prosper since it more easily aligns and provides value in specific functions.  I know of a small specialty chemicals company whose freight department uses analytics to optimize the movement of large amounts of chemicals throughout their supply chain. 

But the company's overall competitive advantage is material technology (making new chemicals to meet customer demand and trump the competitors).  So in this case, the analytics team housed in the freight department supports functional objectives by managing freight in the best way.  But, changing the analytics team from the freight department and asking it to report to the CEO and calling it a CoE wouldn't make sense.

In my view, analytics is best wherever it can best align and support key organizational goals.


Re: By any other name
  • 8/8/2012 4:13:58 PM

Hi Gerry, thanks for clarifying. Sounds like the company has embraced a culture of analytics. Is that fair to say?

Re: By any other name
  • 8/8/2012 11:34:12 AM

Hey Beth,


I am saying that our industry (in its infancy) has a tremendous number of questions and concerns and very few answers.  We are integrating remote medical devices in order to minimize the cost of medical care, increase the productivity of a stagnet number of medical personnel and a balloning number of patients entering into the intense medical care stage of their lives.

In order to make sense of what we are doing - and to assist in making life changing decisions, AET is forced to take calculated risks to guide our healthcare market.  Using the power of analytics based on the massive amount of data that we continue to grow, we need the best available decision analytics.  We feel comfortable in our advice due to the power and of the use of decison analytics.  We have not developed a CoE (as such) but have a group that is responsible for our decisions based on analytics - thus decision analytics.  



Re: By any other name
  • 8/8/2012 10:34:49 AM

Gerry, are you saying that Agile Edge has established an internal CoE and that you call that group Decision Analytics -- or are you saying the company applies decision analytics as part of its systems integration work?

Re: By any other name
  • 8/8/2012 9:49:48 AM



I applaud your use of Decision Analytics - something that we have used for a couple of years at Agile Edge Technologies to identify the real need of analytics in the area of remote medical devices.  AET is an system integrator (device integrator) for the home healthcare markets and it is essential that we bring together the best mix of remote medical devices used in the care of home bound patients.  The use of remote medical devices is an essential part of future patient care.  The more data we collect from the patient (non-HIPPA and SOX restricted) the better our predictive modeling can become.  Since we evaluate future (early stage) medical devices - it is important for us to be able to predict with accuracy the positive effect a device will have on selected patient populations.  We need to know that a product makes a difference in patient care, provides the doctor with access to critical health data and manufacturer with reliable performance data.  At all levels of the medical community - decisions need to be made that can be life-saving.  Decision analytics can made the difference between a poor decision and a good decision.  

Thanks again for your input.  




Re: By any other name
  • 8/6/2012 10:57:41 AM

David, I think you really capture the problem with the name "Center of Excellence" when you say that it "seems to speak more about you than them." Even when business managers are involved in a CoE initiative, the name does smack of superiority and so I can see why some would find it off-putting. I like Decision Analytics -- not to catchy, but says a lot.

Page 1 / 2   >   >>