When Answering Questions Through Analytics Falls Short
Be it a data model or analytics, framing smart risks helps you make good decisions on when to use data for your decisions
--
Business people from all roles hear of the virtues analytics brings to an organization. The overheard comments are rooted in a certain truth — analytics ushers an ability to effectively guide business operations to profitable marketing and cost-efficient operations.
But there are times when analytics can fall short of lofty business expectations. The key to identifying an imminent downfall lies in understanding why analytics should be applied to a given business challenge.
Analyzing data to form a model is not a “snap your fingers” event — despite tools that have “seemingly” made the tasks in creating a model more reliable and easier.
You must rate the importance of the questions you need to solve.
You must know when you’re dealing with decisions that solve good problems or threatening problems.
Not distinguishing the quality of a problem ruins your priorities, because some questions create more tasks — and thus take more effort and time to solve. You want to work on good problems with data models or analytics while addressing threatening problems with a different set of tasks.
· A threatening problem is a situation that poses an immediate danger to the potential success of a strategy or even the existence of your business. It needs a quick response, so you have to know if the data and process can deliver in the timeframe given.
· Good problems are just as important as the threatening ones, but differ by having some time to redress. You want to take advantage of that time to refine that solution. This lets you iterate on the data, such as that in a machine learning process, plus allow time to assess results and develop the resources necessary to solve the problem.
Deciding the problem you have impacts the decision to build a data model. Is it a smart risk that can be quantified to make a decision?
Many models require time to develop. Moreover, ideas derived from data often trigger discussions — a lot of…
