As enterprises sift through their big data, prepare it, and formulate query and analytics approaches, it isn't too early to also start work on a reporting architecture that can present big data in ways that make it most immediately actionable.
In the old days, systems analysts sat down with business users to define online and batch reporting requirements that included report formats, labels for columns, which data appeared in each column, and so on. Big data presents a different set of reporting challenges when it comes to defining reports because its purposes are much more diverse; also, companies are confronted with myriad big data reporting choices from vendors, and that can add to the confusion when it's time to standardize reporting.
The bottom line is big data reports must be actionable. This means that the reporting "forms" you pour big data into must be tuned to deliver results in the most actionable format possible for the purposes for which they are intended.
Presently, big data reports in enterprises come in several flavors:
- "Bottom line" reports from big data that come in the form of system alerts and visual dashboards;
- Graphical summaries of big data that can be further probed for detail; and
- Standard spreadsheet formats that present a combination of data summary and detail information in a columnar format that can readily be manipulated by moving columns, changing sorts, and adding data operands to produce offshoot scenarios of the original baseline report.
All of these reporting formats are capable of delivering great value, but they can only be optimized if a business analyst puts on an engineer's hat and thinks about form, fit, and function.
Here are four examples that demonstrate my point.
- You're the head of e-commerce and want to know how the online store's transaction system is performing right now; do you want a detailed report from your big data to sift through, or a dashboard that flashes red, green or yellow based upon system operating status?
- You're in charge of a city tram system and a set of track goes down; do you prefer an immediate system alert of the problem, or a line-by-line display of status on your desktop or mobile display that you have to read through?
- You're the vice president of sales and want to look at your percent of revenues this year in Germany, Japan, Brazil, and the U.S.; do you prefer a graphical summary with some drill-down capability, or a detailed spreadsheet where you can manipulate columns, but you have to put in extra time that you don't have?
- You work in accounting, and you want to investigate how to better manage expenses to improve margin; are you better served with a graphical summary of data or a detailed spreadsheet with a full set of manipulative tools that can really get you down deep into the data for purpose of discovery?
In each case, you'd want reporting that is immediately actionable, though the types of actions you'd consider actionable vary as broadly based on job function. For this reason, when you go to user meetings to discuss the best way to harness big data to help their business, be sure to find the functional report that best suits their needs based upon how they will use the report.