We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.

Business Analytics for Troubled Times

Originally published March 10, 2009

Everyone realizes that we live and work in a different world today than that of just one year ago. The year 2008 was a turning point in business and in society. Confidence is down, uncertaintyabounds and uncomfortable words – turbulence, recession, bailout, etc. – dominate the news. The economy is the driving force, but it is not the only change. Customer behaviors aredifferent, employee expectations have changed, corporate strategies are fluid and futures are uncertain. When the recovery occurs, it is unlikely that we’ll simply return to business asusual.

So much has changed in the business world, yet strangely one thing remains the same. Many of today’s business managers rely on the same analytics – the same metrics, scorecards anddashboards – as they used in the past. I find this puzzling because the relevance, usefulness and value of yesterday’s analytics have diminished severely. Is analytic inertia –doing it the way we’ve always done it – part of the problem? Does anyone really believe that the analytics and actions of the past will be good enough for the coming months? Are we simplydoing what is comfortable and familiar? Or are we stuck in analytic inertia because we don’t know what we need to do differently?

I believe that inertia is much of the problem and that it results from uncertainty. We do nothing. We remain inert because we don’t know what to do. To overcome analytic inertia we must:

  • Realize that decision making has different implications in today’s business climate than that of a year ago.

  • Understand how and why today’s business questions are radically different than those of the past.

  • Recognize that having different questions implies getting different answers – the metrics change too.

  • Use new analytic processes that are a good fit for the new kinds of questions that need to be answered.

  • Find the right technologies to enable new analytic processes.

The Decisions

Remember when the difference between a good decision and an adequate decision was simply how much growth would occur? In an expanding economy, it was easy to look good even with mediocrebusiness decisions. Boom times! Upward trends everywhere! What a lot of fun! And the few who promoted principles of decision process quality were often viewed as contrarians, academics or simply outof touch with reality.

But events of the past year remind us all that unchecked expansion can’t continue endlessly. The near-guarantee of growth has disappeared, to be replaced by a new reality – decisionquality matters. Forget about bouncing back from bad decisions; they will put you out of business. Also gone is the idea that you can squeeze enough growth from mediocre decisions to declare yourselfa winner. Mediocrity is just a slower path to going out of business. Even good decisions may not be good enough; they buy time, but not certainty. In these difficult times you’ll need somegreat decisions to restore confidence, conquer uncertainty and ensure the survival of your business.

Great decisions – it sounds like setting the bar at an exceptionally high level. Maybe so, but today’s businesses need great leaders who can make great decisions. Those leaders will becharacterized by two things – understanding the nature of uncertainty, and asking the right questions.

The Questions

It is almost cliché to say that these are uncertain times. But look a bit deeper than the superficial meaning and you’ll find something significant. Uncertainty is at the core ofcurrent economic events. Uncertainty is paralyzing. With it comes indecision, lack of confidence and failure to take action.

Simply defined, uncertainty is all of the things that you don’t know. In the myriad of unknown things, some are more important than others. Some are critical to the decision processes whileothers are trivial. Critical uncertainty is all of the things that you don’t know and that you need to know to arrive at high-quality decisions. Critical uncertainty is the key to asking theright questions, and asking the right questions is the key to analytic value.

Let’s compare questions – past and present. In the growth economy the questions were, not surprisingly, growth-oriented: What happened? How much of ithappened? Why or what can I do to make it happen more? Monitoring and managing the good stuff is fun and invigorating.

Monitoring and managing in hard times is not nearly as much fun. It is hard work; it is sometimes discouraging; and the rewards are less frequent and less visible. Even the questions lack theoptimistic tone of growth-era questions. Now we need to answer questions such as: What’s the worst that can happen? What is the best that we can hope for? Whatdo we really expect to happen? Why do things happen? And the mother of all business questions – the one that embodies uncertainty: What if…?

The Metrics

When the questions change, obviously the nature of the answers is also subject to change. In business analytics, answers are usually presented as metrics. Consider a simple scorecard for anexample of how metrics might change to answer the pressing questions of these challenging times.

A typical scorecard presents a list of performance indicators together with a small number of attributes for each indicator: the actual value, the target value, variance between target and actual,and the up/down trend of the actual value.  The numbers presented in the scorecard do a pretty good job of answering the questions “whathappened” and “how much.” When a particularly unfavorable or favorable reading is found, the question of “why” is a subjectof deeper analysis using tools such as OLAP.  Or perhaps it is a subject of conjecture that is answered by gut-feel, as suggested at CIO.com in an article titled To Hell with BusinessIntelligence.

In this new and more difficult business climate, the questions are harder and the metrics need to be more robust. Four simple elements – actual, target, variance, and trend – no longer dothe job. By expanding the metric to eight elements,
  • Actual value

  • Directional trend of the value

  • Expected value

  • Variance from expected value

  • Worst-case value

  • Variance from (proximity to) worst-case value

  • Best-case value

  • Variance from (proximity to) best-case value
I now have the information to answer some of the new, tougher questions: What’s the worst that can happen? What is the best that we can hope for? What do wereally expect to see? And I have pretty good indicators to change expectations – to revise either up or down – based on the reality of what is happening.

This is progress; now we’re moving in the right direction – richer metrics providing deeper answers to harder questions. But the big questions that are the core of critical uncertaintystill remain: Why do things happen? What if …?

The Analytic Processes

The most common analytic processes – those found in most business intelligence (BI) programs – use a sequence of goal-setting, measurement and monitoring. The monitoring activitysupports a rudimentary feedback system where the essence of the feedback is “on target” or “not on target.” This process is sufficient to manage in a growth economy where thequestions center on what and how much.

Today’s harder questions – why and what if – demand a more advanced analytic process. We need to shiftfrom a process of goals-measures-monitoring to one of modeling-simulation-feedback.

Modeling supports the need to know why things happen. We need to understand cause and effect – to understand the influences of things upon other things. Knowing why means knowing what leverscan be pulled to effect change. An advanced analytic process needs to include cause-and-effect models that capture the knowledge of influences and the complexities of why things happen. Causal loopdiagramming from the systems-thinking community is a good fit here. You can read more about causal modeling in the articles A Systems View of Business Analytics, Part I – Introduction to Systems Thinking and A Systems View of Business Analytics, Part 2 – Recurring Patterns in Systems.

Simulation goes beyond why to answer the what-if questions. To simulate requires more detailed causal models than thosedescribed above. Simulation seeks not only to understand influences, but to quantify them. We need to know the strength of each influence upon the things being influenced, and we need to know theimmediacy or latency of each influence. The systems-thinking model that supports simulation is known as stock-and-flow. Figure 1 illustrates a simulation model for a manufacturing scenario where theproblem domain is workforce management. The modeling technique is explained in A Systems View ofBusiness Analytics, Part 3 – Making Cause and Effect Measurable.


Stock-and-flow models prepare you to answer those what-if questions by addressing six principles of simulation:
  • You know and understand the problem domain.

  • You want to influence the behavior of the system, or know how it will respond to external influences.

  • You understand the dynamics of the system – the influences and timing –within the problem domain.

  • You know which system variables are receptive to external influences.

  • You can assign equations to relationships between system variables.

  • Behavior-over-time graphs will tell you what you need to know.
You are now ready to simulate system behaviors – to know “what if … happens.” You can change assumptions, expectations, worst-case, best-case, and many other variablesand observe the outcomes presented as behavior-over-time graphs. Simulation gives you the ability to play with scenarios, to peek into the future and to extend analytics beyond insight to providingforesight.

The Technology

Simulation is clearly something that requires software tools. You can’t build and apply stock-and-flow models with pencil and paper. The technology for this kind of simulation is found inthe systems-thinking community. I am aware of two software products that fit the need. While neither of these vendors positions their products as business intelligence tools, they clearly have a rolein business analytics:
  • iThink (www.iseesystems.com) – Modeling to simulate business processes and scenarios.

  • Vensim (www.vensim.com) – Simulation software for developing, analyzing, and packaging high-quality dynamic feedback models.

Final Thoughts

Simulation doesn’t replace more mainstream analytic technologies – it complements them. But it may be among the most important of analytic technologies. Some knowledge of what thefuture holds will certainly help to apply mainstream analytics in the right places and in the right ways. More importantly, answering questions about why andwhat-if will certainly help to navigate in today’s business and economic conditions.

  • Dave WellsDave Wells

    Dave is actively involved in information management, business management, and the intersection of the two. He provides strategic consulting, mentoring, and guidance for business intelligence, performance management, and business analytics programs.

Recent articles by Dave Wells



Want to post a comment? Login or become a member today!

Be the first to comment!