Strategic BI Considerations: Enhance your BI IQ series
Recent research from Stanford and Princeton, confirms Gartner’s recommendations ‘Without Business in business intelligence, BI is dead’. For, when you leave you BI projects solely to your System Integrator/s their possible path may become delivering a large quantity of reports and analytics. While what your business might need is high quality analytics with self service capabilities.
Recent research clearly demonstrates that though this, large number of reports, seems alluring the path could lead to lower quality decisions.
When a business user needs to make a critical decision there are three scenarios they normally go through
1. The Good: they know exactly what report they need to access to get the required information. Or, they know how to build their analytics with modern ‘Self Service’ reporting applications
2. The Bad: They have over 40 reports that may contain the information they need and they will probably need to go through tens of reports to find it any of them or all of them contain the information they need. some of these reports may be junk reports. Such a situation leads to further confusion if each of the reports gives numbers than contradict other reports.
3. The Ugly: They have hundreds of reports in the BI environment but most of them cannot be used by business to take any kind of decision, leave aside critical decisions. The SI came, delivered hundreds of reports and they left by week 2, leaving business users hyperventilating with unusable reports and analytics.
According to Gartner’s 2012 report over 70% of BI initiatives live in scenario 3; less than 30% live in scenarios 2 and 1. You need to conduct a thorough ‘Strategic BI Health-check’ to see where you actually are. Automated solutions are available to accomplish this internally.
In almost every situation where users are subjected to information overload they tend to make worse decisions, according to a fascinating and new research from Stanford and Princeton psychologists.
The classic study from Anthony Bastardi of Stanford and Eldar Shafir of Princeton presented research subjects with scenarios in which they needed to choose between one of two alternatives. The example was totally related to the students as it dealt with credit applications for student loans for a course they wanted to take.
In these experiments sometime all the required information was supplied and at other times key information (for the required decision) was eliminated from the information provided. However, after the decision was taken these key information elements were exposed to the respondents. So from an emperical point of view all the respondents got exactly the same information but with one key difference. In the first group all the information was provided at the same time. With the second group there was a delay to providing the key information elements.
The result was astonishing from one side, but quite predictable from another. The question being asked was whether both groups finally took the right decision. The answer – Not at all.
It was clearly established that the group that had to wait for the key information elements at a later stage they consistently said ‘No’ to the course they wanted to take and refused loans at a higher rate.
So what is the business impact for a information consumer, or a decision maker, in an enterprise. What is the lesson to be learned?
It clearly establishes that when humans are made to wait for information, or if it is not readily available they assume it to be more valuable, even if the information bits are not relevant or important. So decision makers get into a ‘freeze’ moment if they assume their information is incomplete as they then believe the missing information is the most critical bit required to take their decision.
The Human mind is wired to instinctively collate unstructured information in situations and make logical sense out of it. When we hear a sound in the forest we can pretty accurately guess if it is make by a human, a squirrel or a potential threat like a cougar. We are wired to logically reduce uncertainty during hazardous times.
Howevr, in enterprise decisions there is a subtle difference. The first is that in todays flat world each decision could have tremendous impact on the strategic health of any company, i.e. a tiger in every bush scenario. Add to this the fact that we instinctively add more value to data that is missing. There is an inherent counter-benefit need to look for additional information- sometimes even if the information provided is adequate for taking the right decision.
In such cases if there are 40 additional reports that also deal with what sounds like similar information the user will be inclined to review all these reports hoping to make a better decision. However, if some of these reports are ‘junk’ or ‘unusable’ reports these will lead to add to the confusion or non-decision attributes of thinking. If this ‘junk’ number is high it can potentially freeze enterprise decision capabilities.
In the experiment ever the group that was provided with all the relevant information could not help but ask for more information. In enterprises where such additional information is available they become forced to go through it all and often end up more confused that when they started.
In a badly designed BI environment there are unused, unnecessary and test reports that sit in production BI environments, i.e. the ones business actually uses. The mind of most decision makers hates information gaps and often their attention gets focused on the wrong information elements. For example as the respondents got confused their attention got focused on the loan amount , i.e. 5,000 or 25,000 and not on the most important aspect that the student had a history of defaulting on their loans. In such a situation the loan amount was a minor detail.
If your user, in their quest to fill the information gaps, starts to click on additional reports, which by definition are ‘junk’ reports they will only end up more confused as some of these reports may hijack their mind to inconsequential details that do not really matter.
Reality is that users will want to read just one more report, one more analysis which today is only a click away. But if such reports are not harmonized for data and information consistency it may actually degrade the decision capabilities by clouding their information assets, or worse actually make them loose confidence in the BI reports.
So next time you want to make a good BI make sure it does not only deliver quantity, but quality
The next time you want to make a good decision make sure you spend more time in defining exactly what type of information do you need to take the decision and focusing on just that.
The next time you go into your BI environment and find ‘Junk’ reports, put into place a process to eliminate them all from your production envirionment.
Don’t trip yourself with junk reports. Don’t trip yourself with redundant analytics and most importantly of all don’t trip yourself with non-value-add information.
Lesson: 'Conduct a 'Stratgic BI Health-Check' starting next monday.
Question: How do I start a process by Next Monday: to find what reports in my BI environment are 100% usable in my DI environment. Read BI Valuenomics