Just because you make beautiful, impactful data visualizations doesn’t mean the business is using them to their fullest extent. There are ways to improve that.
Image: Mykyta Dolmatov, Getty Images/iStockphoto
A familiar story we hear in big data management is the 80:20 rule: It takes 80% of data scientist (or IT) time to clean and prepare data before it can even be used for a report, then the other 20% is spent analyzing the data. But there is another 80:20 truism that organizations need to start looking at. Big data reports, once developed, are likely to fall into the same conundrum as traditional IT reports: Only 20% of the reports will be actively used, while the other 80% are seldom or never used.
SEE: Report: SMB’s unprepared to tackle data privacy (TechRepublic Premium)
Here are some challenges for big data reports:
Lack of business relevance
Table of Contents
Most organizations work hard to define their business cases for big data reports, but the nature of big data and algorithm testing, which is repetitive and ongoing, is that “drift” begins to occur. In other words, while you are on the way to capturing data and insights for a specific business case, other side-issue questions also begin to arise. This can be healthy: It means you are exploring many different uses for your data. However, the danger is that so much drift occurs that you forget what your original business case mission was. Consequently, you end up missing what you had hoped to achieve, and the results of your efforts are diminished.
User illiteracy
IT and many vendors provide a set of predefined big data reports and dashboards that are meant to assist different areas of the business, such as finance or marketing. But just because these reports are deemed to be “best in class” doesn’t mean that your users will understand how to get the most out of them. In some cases, the terminology of the reports and dashboards might be unfamiliar. In other cases, the metrics and information that the reports and dashboards display are not ways that your business has chosen to measure its progress. If users don’t understand what they’re looking at and can’t see how the reports and dashboards solve business problems faster than users can on their own, the reports won’t be used.
SEE: How to create your first data visualization dashboard in Tableau (TechRepublic)
Overwhelming data
Sometimes dashboards and big data reports are so rich with information that they are simply overwhelming for users. This is where less really can be more and where a reduction of options and data can help users focus on the primary issues they need to solve.
How to guarantee your big data reports get used
Those heading up big data and analytics functions in companies can do several things to ensure that their users are getting the most out of big data dashboards and reports. Here are three recommendations:
1. Stick to the business case
Once a business use case is determined for a big data dashboard and reporting project, stay with the specification! If side issues arise, put these into a “parking lot” of ideas that you can address later.
2. Maintain a close collaboration with users
Your big data reports and dashboards will be successful to the extent that users use them. The best way to ensure that dashboards and reports are relevant is to listen to user needs and periodically check in with users to see if any upgrades or changes need to be made.
3. Treat big data and dashboard reporting like any other IT reporting
A standard IT practice is to review reports that users receive on an annual basis and determine the extent that each report is used. If reports are seldom or not being used, get together with users to see if the report should be modified or discontinued. What you don’t want is a warehouse full of shelfware that no one wants.
SEE: Develop a Big Data reporting strategy for all users (TechRepublic)
This also goes for big data reporting and dashboards, which should be subjected to the same level of scrutiny that older IT reports are. If big data reports and dashboards aren’t being used, discuss them with users to see if they can be modified. If they can’t be upgraded for use, discard them.
Data, Analytics and AI Newsletter
Learn the latest news and best practices about data science, big data analytics, and artificial intelligence.
Delivered Mondays