Reporting and Analytics

I’ve embedded this topic in several times in different articles, but I’ll write it again specifically just Reporting and Analytics. Because I came across it again yesterday and it is very important concept in BI. As a user, and equally important, as a BI architect.

As a user, you would want to do 2 things in BI: reporting and analytics.

1. Reporting is where you read a report. Columnar tables. Layout. Appearance, corporate logo, cosmetics. Regulatory (to get certain information at certain time for a certain government body). Font, header, footer. You know what information you need. You need a report built for you to get the data out (of the warehouse, and of the ODS, and of the transaction system). So you have warehouse reports vs operational reports. They layout is very important to you. It has to contain specific columns, etc.
2. Analytics is where you explore a cube. Slice and dice. Browsing the data. Drill down, drill up, drill across. You have an idea of what you are looking, but you don’t know the specifics yet. For example, you want to find the trend of premium and claim reserve across different classes but you don’t know which measures and for which period. You want all the data in the system at your finger tip. You want to explore the data to find things. Things that you don’t know yet for sure, but the further you explore the data, the clearer the picture in your head. Then after you get a clear picture, you want a dashboard. Things you need to look every day. You want KPI, your team’s targets. You want BPM, tracking your budgets.

Years ago in many companies where ERP was introduced there’s a need of reporting. Users ordering reports from “the IT department”. 4 GL reports, Easytrieve, Cobol, AS/400 reports. Everything is on Mainframe. Then mini computers. Then everything is on PC. R&R was very popular. Clipper. FoxPro. Access. Suddenly everything became PC based. Crystal Reports was number 1. Still, the essence is the smae: users ask IT to extract the information via reports. The “report developer” profession at that time was very popular.

Then the BI era. Dimensional modelling. OLAP. Cognos, Essbase/Hyperion. MicroStrategy. Cartesis. Concept of dashboard, KPI, score cards. BPM/EPM (Hyperion was right in the front). The whole world was doing BI & DW. Until today. But the old reporting still exists. Not as elegant as before, but still required. Strongly required. For routine, monthly progress report (sales, production, etc). For regulatory purpose (FSA, etc). For the Senior Management. Operational reports for the sales managers, branch managers. Reports still exists strongly today. But it never be a selling point when a BI consultant pitches their products & services. Usually it’s analytics and visualisation. They are always the key selling points. Secondary selling points are DQ, MDM, DG (data governance), DI (data integration) and the “realtime-ness”.

So if someone asks do we need reporting or analytics, the answer is clearly both. Reporting is the basic of “information extract”. Users can survive with reporting alone. But not for long. Your company can’t excel or perform without “BI”, i.e. the analytics. In my book I said there are 3 aspects of BI: reporting, analytics and data mining. Forget about mining for now (predictive analytics, etc). Let’s focus on the first two. This article is about the first two, reporting and analytics. You need both. Organic growth of “information extract” within a company will start with reporting, but then moving onto “using that extract intelligently”, “exploration”, “trend analysis”, etc and before you know it the company is in the middle of BI. Dashboard, KPI, BPM. BI has major selling points (from the IT director to the board). Hence it’s taking off everywhere in the last 5 years. Yes the economic downturn in the last 3 years killed off many BI projects, but it’s still one of the most important ones on the agenda of CIOs. Even with limited budget, some BI projects still need to continue. Tactical ones especially. Within the targeted business area, usually the one that is a key revenue generator to the business, i.e. credit (bank), underwriting (insurance), sales (retail), tariff (telco).