Deloitte on the Business Value of Visual Analytics

Several Deloitte partners have written an excellent article in the Wall Street Journal emphasizing the business value of using visual analytics to help managers reduce costs, identify anomalies and exceptions, and develop a better understanding of operations that can inform strategic decisions.

They point out that the typical visualizations most people see in dashboards, presentations, and reports — pie charts and their cousins — can be confusing at best and misleading at worst. It’s simply much too easy to hide useful information in aggregated data, especially when you present that data in a visual form.

The authors elaborate on three main points:

  • Ask the right questions
  • Transform the data
  • Apply visual analytics

The full article is well worth reading, but for now let’s look in a bit more detail at one often underappreciated issue: Asking the right questions. They point out, “Understanding what questions the organization is trying to answer enables the development of more effective and efficient cost and profitability models. The process of defining questions can also provide a guide for visual analytics and data analysis overall.”

I agree completely. It’s incredibly easy to “see what you want (or expect) to see” when you ask only the most obvious questions. Good managers need more than access to lots of data; they need to be able to manipulate it, filter it, and visualize it in truly effective ways in order to make good, informed decisions. The writers bring up these examples of good questions for a large enterprise in the article:

  • What’s driving profit performance, and what areas in the business need attention?
  • What are the levers to reduce overhead and shared services costs?
  • What’s the total cost to serve by customer, channel or region?
  • What’s driving swings in margins?
  • How is product mix impacting the business?

In Capital Markets, we might ask more targeted questions like these:

  • Which execution venues are performing best – and worst – for a particular instrument?
  • Which traders are underperforming the market and/or their peers today? During this hour? Right now?
  • Do we have any algos that are underperforming today compared to the past few days?
  • Are we seeing a large number of small, unprofitable trades that are undermining our overall performance today?
  • How does our trading activity right now compare to our activity from earlier today in terms of profitability, execution times, best execution, and exposures?

If your people have a solid grasp of the questions that need to be answered on an on-going, real-time basis as well as at end-of-day or end-or-period, it makes configuring your data infrastructure and designing your analytical dashboards almost (not quite, but almost) trivial — at least in terms of the conceptual work involved.

Read the full article here: