It was an amazing game. The New England Patriots defeated the Atlanta Falcons 34–28 in an historic overtime comeback. Why were they the best team on that day? It was because they had the best specialists. If they had allowed any player to fill whatever position was in need, they never would have won. For example, Tom Brady is arguably the best quarterback of all time, but if he had been playing defensive lineman, would the Patriots have won? Likely not.
Analytics teams are no different. Companies with the most impactful analytics teams are generally the ones with the best analytics specialists.
Then why is it that most analytical teams I have encountered don’t develop specialists? Why is it that many of them have only jacks-of-all-trades, who are tasked with doing many things: building reports, answering executives’ questions, conducting intense R&D and working with IT to operationalize analytics.
To assemble your company’s analytics dream team you’ll need deeper expertise, likely including the following specialists.
These folks create tools that can be used by many individuals in the organization to understand the data. By creating static or dynamic reports along with dashboards or exception reporting, they enable the organization to make decisions using well-defined and reliable information. Fact-finders must be experts at BI tools like Tableau or SAP BusinessObjects, and understand production data structures and availability. Generally, they will only access fully prepared data.
Typically, reports are designed to answer questions that begin with “what.” For example, “What were the sales in California last week?” or “How much inventory do I have in the warehouse?”
Storytellers conduct ad-hoc analytics to answer fast-paced questions from business P&L owners. Typically, questions begin with “why,” such as, “Why did California sales decline last week?” or “Why are we running low on inventory?” Successful storytellers can piece together information from many different sources, and create very compelling answers to these questions.
Not only will storytellers use existing reports, they may also extract data from multiple data sources to best tell their stories. They need to be fluent in SQL, as well as other tools that allow quick merging and aggregating of data. The most productive individuals also have training in advanced analytics modeling, such as statistics or machine learning.
Explorers answer very complex, forward-looking business questions, unlike fact-finders and storytellers, who are largely looking at the data historically. Their questions begin with “Can we,” such as “Can we predict future demand to improve our supply chain replenishment?” or “Can we determine the next best offer for each of our customers?” Think of explorers as analytics R&D, where there are no guarantees that every analytics initiative will result in business-changing analytics findings and insights.
People that are successful in this role are true data scientists—very adept at using tools such as R, Python, Matlab, SAS or Aster. They have very deep analytical skills, and are proficient at statistical or machine-learning methods.
Architects design the solutions and work with the storytellers and explorers to ensure that the designs are implemented correctly. Who in your organization does this? Unfortunately, this task is often left to the story-tellers or explorers, but they aren’t architects.
To fill the gap, organizations typically pursue one of two unproductive solutions:
- Analytic teams execute many operationalized solutions themselves, avoiding the time-consuming, and often politically-painful process of implementing rigorously and properly with IT. I call these cowboy processes, and while they are not always bad, they can create substantial organizational risk and should be tracked and understood.
- Or, they have analytics professionals write the requirements, and then provide them to the developers to operationalize the analytics. This is a recipe for disaster, as analytics professionals don’t typically communicate in IT-friendly ways and, ultimately, the result is often poorly designed or incorrectly implemented solutions.
Ideally, organizations identify architects who specialize in working with IT developers, who understand both the analytics and the culture of IT. Their job is to work closely with the analytics, business and IT departments to ensure the analytics are implemented rapidly, correctly, and make sure any IT concerns are well-understood by the business.
These folks act as a liaison between analytics and business. They are incredibly aware of business objectives, and, as a result, have very high business acumen. They also understand analytics, can explain complex analytics in business-friendly terms and can apply the analytics to show P&L impact.
I realize that in smaller organizations, it is not practical to hire individuals who have a single role. In those cases, I challenge you to give each individual only 2–3 well-defined roles, to ensure as much efficiency as possible.
Your competitors are investing in analytics as well, but building a Super Bowl-caliber team of specialists will give you a significant advantage (even if you find yourself way behind, and needing a last-minute, game-changing comeback).
Scott Langfeldt is a Senior Analytic Consultant for Teradata. His 20 years of data science experience designing and implementing analytical business process solutions has given him an appreciation of the benefit of analytics on organizational performance and how to effectively implement these solutions in production processes. In addition, he brings an understanding and passion for identifying and preventing obstacles for analytic excellence.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise.