»Me, myself and BI«

Bissantz ponders


More numbers! More instinct!

On May 27th, we will discuss the topic ‘Management between Intuition and Automation’ at the Hotel de Rome in Berlin. Human intuition filters and summarizes details into signals. Business Intelligence wants to do that as well. Does management need more intuition? Does intuition need more numbers?

This May, we will be heading to Berlin again where we will discuss the future of Business Intelligence – just as we already did in 2009 und 2007. The motto of this year’s event is „Management between Intuition and Automation“. This time, the extreme mountain climber Reinhold Messner will join us and talk about risk management. As he says, “I don’t dare go against my instincts.”

Once again, there will be plenty of room for new ideas in the ballroom of the Hotel de Rome. In days gone by, bank managers could peer from their offices into the former service counter area. This view of the happenings around them complemented their factual knowledge with intuitive impressions.

We use our instinct and intuition when we need to make a fast decision under uncertainty and incomplete information. In fact, this always occurs when a manager makes a decision because at least one – if not all – of these factors apply. But wait, isn’t Business Intelligence supposed to correct this situation…through signals, aggregation and filters?

  • Signals I’ve often moaned about the dangers of using signals for management . This message, however, is worth repeating. Whoever relies on the signals of others will be let down quickly. I completely agree that one of the causes of financial crisis is that people blindly accepted the AAAs and BBBs of rating agencies. We’d better find signals for our actions ourselves – by closely studying relationships and details.
  • Aggregation According to the fans of dashboards and scorecards, details are something that we should keep away from managers because they simply overwhelm them and enough of the signal remains in an aggregated form. Of course, profit and loss and a balance sheet are a good representation of the big picture and take time to read in detail. But what do managers do the rest of the day? They pay attention to details without the help of reports. They walk through the company and speak with suppliers, customers, employees or investors. Just a few days ago, an executive friend of mine commented how much attention he has to give to find the truth beyond the numbers. Every aggregated performance indicator related to output, he feels, is irrelevant as long as he can’t see if the produced quantities fit in the structure of the current demand. Time and time again, he also sees that aggregation sows the seeds for manipulation.
  • Filters Nevertheless, we still need aggregation and selection in some form. That’s why we wish we could have intelligent algorithms that detect the details we want to know as a whole. The human ability to recognize patterns sets a pretty high standard. In our own company, we have a control center that spurts out over 1,500 memos, visitor reports, protocols, progress reports, support cases and other texts from our CRM system. Our management team only takes a closer look at a few of these but since the human brain is so good at making associations it is relatively easy to piece them together to see the complete picture. Currently, we are working on rules to automatically interpret these entries to increase the throughput of the details. Our goal is to help our management team soak up more of our daily business with the same amount of time and effort. This is a sort of systematic, electronic office grapevine based on data mining. The human eye-brain system can reduce a text to its key messages with unbelievable speed and accuracy. Teaching that to a machine, however, is a very difficult task. Up until now, we only use extremely robust rules for our control center. These, in turn, are based on domain knowledge that is nontransferable and specific to our company. We have already invested 200 hours in our thesaurus alone.

As we can see, good BI shows many details – and, more importantly, the right ones and in the right format. Automation helps us with this task. The computer can take over search processes, build hypotheses for large data pools, create patterns, support an integer presentation based on rules, generate suspicious facts, and imitate intuition in cases where routine would simply take too much of our time. We thereby support and in parts replace intuition through automation.

That takes us back to our event in Berlin and Reinhold Messner. Messner stands for extreme decision-making situations. In his expeditions, he regularly faced life-and-death situations. The utter lack of paths that Messner faces in the most forbidding parts of the world goes far beyond of the imagination of us Alpine novices. To assess the situation, he had barely more than his own intuition, perhaps a weather forecast and the expert opinions of his partners. Managers, in turn, have many aids to support their decisions. Can – or must – managers try to go against their instincts? Aren’t instincts simply the sum of all conscious and unconscious facts, filtered through experience, concentrated on emotional impulses and calculated in our heads as quick as a flash? If so, it is the job of Business Intelligence to help – specifically, to support the modern view of management and ultimately bid farewell to the one-dimensional monotony of a simplified style of management by performance indicators.

I am looking forward to Berlin, Reinhold Messner and the effect that this event will have on us and others as it did twice before.

    Leave a response

    »Me, myself and BI« Bissantz ponders