How to define quality business intelligence
Doctors face it, engineers face it, insurance companies face it and students face it, that little voice in the back of ones head that raises a small question or doubt. Sometimes we ask for validation on information provided and discover that the what we thought of fact y turned into fact z. It is the “Do you know what you think you know?” question.
Quality is defined as fit for purpose. Therefore in some circumstances it can feel rather subjective. A muddy glass of water in the jungles of Borneo may taste as good as a fresh bottle of mineral water to an exhausted traveler. However if we place the same traveler back at home in Madrid, they may refuse to consume it.
For the same reasons and understood by those readers of George Orwell, we know that not all business intelligence reports are made equal. There is a vital role in data collection to ensure quality of data capture is fit for purpose. That definition may need to change as a business, market or organization evolves and so keeping abreast of such changes is an important resource decision by leadership.
In this world of funky info-graphics, 3D charts and ever more creative visualization tools to represent data, we need to remind ourselves of the basic backbones of this data. As leaders it is important to remain positive and yet critical of the data supplied to define its accuracy.
An open culture in data presentation can open the eyes and increase the speed of business performance. How? Well, let us compare two scenario’s.
1) A production manager shows a graph with the average number of widgets produced in 2 hour time period. This allows the sales director to calculate the number of widgets he/she can expect to deliver in the next week.
2) The same production manager shows the same graph but states the accuracy of the data has an error margin of 3-5%. This allows the sales director to calculate a much more realistic number of widgets to be delivered over the next week and thus satisfy the customers with more certainty.
OK; this is a simple scenario but you get the picture, as more and more complex data sets are created the conclusions made from them become ever more prone to error or higher risk consequences. So, by being open within the business culture to accept data error bands then a business can become more prepared to pivot and gain competitive advantage as the reality of performance is realized.
The quality of your data and data capture systems will define the difference for your business between operating business intelligence system (BI) as opposed to a business stupid system (BS). We all know that BS does not stand for very long!
For this purpose I suggest that leaders question their questions, question the quality of data being provided and allow the foundations of your business intelligence systems to become robust. Once this sound strong foundation is in place the delights and fun and advantages of data mining, accurate OLAP (online analytical processing) provide the back bone to great commercial decisions and when you are ready to engage in the world of big data, your own quality contributes to a more trustworthy and beneficial system.
Check that your business intelligence system stakeholders have the skills to operate it.
Check that your stakeholders also have the correct time to use the business intelligence tools supplied.
Expand your data collection beyond basic excel reporting.
Before expanding to a data warehouse define your business intelligence wish,wants and needs.
Define your business budget for business intelligence for the next financial year.
Before expanding your business questions, expand your business data quality.
Did you copy,paste and action the above steps? No? Well, try it, you got this far and I would not enjoy sharing this information if you did not use it! Indeed I prefer that you stop reading right now and action the above steps before you continue.
So, did you? OK I’m kidding, now that you have the above process underway you can march forward with increased confidence in your business intelligence quality. You now have some defined steps and awareness of how and why it is important to question and drill into the data shared within your business. Whilst the business adjusts these areas and establishes the changes it is clear some parts of the business will often be faster at implementing that others. Therefore you can encourage this positive change management direction and reinforce the actions with an extra little tip.
For the next months round of information presentations ask the presenters to “confidence color code” their slides of metrics and reports. Green, yellow or other suitable colors will help adjust the mindset of the management team in knowing the leaderships is aware of in accuracy within its reporting system. The leadership can embrace and celebrate the movement by seeing how many reports are displayed in any relevant color. The change to attitudes in reporting can offer you new insights to your operations.
Leaders and stakeholders within the whole business intelligence system need to remain realistic. It does take time and resource, so the benefits of system improvement need to be measurable to a direct benefit back to the organization. Therefore a 10% error in some reports may be OK and cost-effective, in other areas however it will become critical to even have a 1% error. So, make your judgement call and be aware of its consequences or benefits.
Further reference material
10minute YouTube video : Todd Schwartzrock presents the case study of Chrysler’s customer data warehouse, a huge success in data quality and continuous quality measurement. This program has achieved a 98% quality levels on over 50 million records.
Also this pdf document via the 16th International conference on information quality served from 2 Australian Universities
Right now take a short moment to scroll down and share this article to anyone you think will find it relevant and interesting.