I recently read an article from Software Advice called the BI Buzzword Breakdown. The article highlighted the confusion around business intelligence (BI) terms and asked experts to provide their own definitions. While a quick visit to wikipedia will define most business intelligence terms, we as practitioners have learned that successful BI implementations require significant support and interaction with business users. The IT-led 'build it and they will come' strategy, just plain does not work. Thus, we need to be able to translate from techy terms to business terms. As a compliment to the software advice article, I would like to attempt to explain some of the key terms as if I was speaking to a fictitious business executive, who I will call Amy.
Amy is Vice President of Procurement at a global pharmaceutical company. She remembers the days when she got her messages on little pink handwritten slips handwritten from her secretary, typed memorandums for internal communication and carried that little black device that would buzz every time her husband needed to reach her. While Amy has come a long way on the technology curve over the past 10 years (priding herself on her new iPhone), her seniority has allowed her to limit her technology exposure to responding to the endless list of daily emails, creating PowerPoints and occasionally working with Excel. From Amy's perspective, she is focused on managing the procurement process and gets most of the information she needs (or thinks she needs).
Last week Amy joined a colleague at a procurement conference and heard a bunch of Business Intelligence terms and has asked me to explain them to her. So I gave it a try. Here is a summary of what I said . . .
Amy, thank you for inviting me in today to explain the basic concepts related to business intelligence. Given that your focus is on procurement, I will attempt to explain the concepts and use examples in a procurement context. Let me begin with a question. Do you receive 'procurement-oriented' reports in a regular basis (e.g., Open POs) and do you analyze those reports to gain insight into the performance of your organization? If so, congratulations!!! You are already doing a basic level of business intelligence. Simply put, Business Intelligence is about using information to gain insight into how an organization is performing.
Now, let's take what you are already doing and go one step further. What if you wanted to see a summary report or bar chart that showed Open PO's by buyer? Or, open PO's buy buyer over the trailing 4 quarters? Those would be more advanced versions of business intelligence commonly referred to as analytics (e.g., when you look at data in aggregate, spread out over a time horizon). Often, companies will put multiple analytical reports together on one screen and call it a dashboard. In summary, traditional reports, analytical reports and dashboards are simply different types of business intelligence.
Interestingly, many of the terms in business intelligence have been around for several decades. While many of the concepts have remained the same, there have been two changes that occurred. First, over the last decade, the cost and capability of the enabling technology. Hence, business intelligence is no longer limited to the leading edge companies with large IT budgets. Second, particularly with the introduction of social media tools like YouTube and Facebook, the volume of data has exploded. No longer to organizations look at internally generated data, they capture and analyze large volumes of data (a.k.a. Big Data) that is being generated outside of their organizations.
Now, you mentioned an interest in understanding of two other key terms, Data Warehouse and ETL. A data warehouse is simply a database that contains a copy of data from one to many of your company's systems. While you can typically report directly upon data in your systems, those systems are intended for processing transactions (e.g., creating and executing PO's). They are not optimal for reporting. Much like you would buy a sports car to drive fast vs. a luxury car for comfort, reporting from data in a data warehouse is much faster and easier than reporting from your existing systems. Further, the data warehouse can easily combine data from multiple systems.
Now that you understand what a data warehouse is and why you need it, lets talk about how data gets into the data warehouse. ETL, which stands for Extract, Translate and Load, is the automated process of pulling data from source systems and correctly organizing it in the data warehouse. Traditionally, ETL is performed once a night, thus the data reported in the data warehouse is only as fresh as the previous night. However, there are many cases where ETL can be executed more frequently and sometimes to the extent where the data in the warehouse is near real time.
I hope that this helps you in having a basic understanding of the terms. Please send me questions if you would like additional terms defined.
