The promise of self-service analytics is almost too good to be true. Business people get the information they want, when they want it, how they want it, while the information technology (IT) department gets to offload report creation to business users so they can focus on broader enterprise tasks. It’s a proverbial win-win situation. So what could possibly go wrong? 

Well, just about everything!

I’ve been studying, writing, and speaking about self-service analytics for almost 20 years. I’ve watched countless organizations struggle to implement self-service analytics. In fact, I’ve seen many instances where self-service analytics exacerbates information access and delivery problems and bottlenecks, rather than solve them.

Nonetheless, self-service analytics is still a goal worth pursuing. Some leading-edge organizations have solved the self-service riddle with a combination of technology and governance controls. And while analytics products today are easier to use than ever and give casual users unprecedented ability to create their own data sets, reports, and dashboards, many now fold these capabilities within a granular permissions framework that curbs excesses of self-service analytics.

Dissecting the Problem 

Self-service analytics can go awry for two reasons: underuse and overuse. With underuse, business users find self-service tools or data too complex and stop using them unless they are given personalized and ongoing training and support. With overuse, business users generate a tsunami of conflicting reports and dashboards that undermine data consistency and create widespread distrust in the data. Both problems often appear simultaneously in a single organization.

Underuse. The irony of self-service analytics, says Kevin Sonsky, senior director of business intelligence (BI) at Citrix Systems, is that it “requires a lot of hand-holding.” Sonsky, like other BI directors, quickly discover that even highly skilled data analysts need one-on-one training and assistance to become proficient with self-service analytic tools.

The poster child of self-service analytics products is Tableau. Although the product makes it easy for users to quickly build visualizations from spreadsheets and local data sources, it takes them quite awhile to master the tool and deliver analyses that answer substantial business questions. Besides formal training, online tutorials, and one-on-one training, data analysts often need ongoing support—an analyst forum or community of interest—that they can turn to with questions about tool functionality and techniques.

More significantly, data analysts can also get quickly over their heads when sourcing complex operational data with cryptic column names, hidden business rules, and table dependencies. Here, the self-service bottleneck moves from creating reports and dashboards to creating data sets. New data preparation tools coupled with online data catalogs and dictionaries promise to eliminate the bottleneck by giving analysts self-service data mash up capabilities.

Things get even worse for casual users. Most forget how to use the self-service tool or are too busy to learn. And many simply retreat into old habits and continue to send requests for custom reports to local analysts or the IT department.

Overuse. At the same time, some data analysts and consumers latch onto to new self-service analytics tools and—like a dam bursting its walls—generate a continuous stream of reports and dashboards that contain conflicting metrics and data that undermine enterprise data consistency.

Acting independently or at the request of a business unit head, these analysts source data directly from operational systems and define metrics, dimensions, hierarchies, and aggregates independently without heeding corporate data standards. These analysts and their business leaders quickly become wedded to their data for tactical or political reasons.

When the CEO convenes a meeting of these individuals, the discussion invariably disintegrates into a dispute about whose numbers are right. At this point, the organization degenerates into data chaos, since no one trusts any data but their own.

This so-called dueling spreadsheet phenomenon drives CEOs bananas. They can’t get answers to simple questions, like “how many customers do we have?” It also frustrates business users who must search through dozens or hundreds of reports to find relevant data. And it exasperates the IT department since data-hungry business analysts often use self-service analytics tools to generate massive queries that cripples the performance of the data warehouse or other systems.

As a result of under- and overuse, many organizations struggle to succeed with self-service analytics. Rather than liberating business users from the shackles of the IT department, it often makes the situation worse. Due to a flood of conflicting reports, business users drown in data while starving for information.  

Towards a Governed Future

The remaining blogs in this series examine how to avoid the pitfalls of underuse and overuse. In particular, they examine how to balance self service analytics with governance controls to provide organizations the best of both worlds: the freedom to access and analyze information quickly and flexibly while maintaining data standards that ensure everyone speaks the same data language.

To start that journey, it’s important to recognize that self-service analytics is not a homogeneous set of capabilities. The next article in the series explains the importance of tailoring self-service functionality to each and every individual in the organization and the various business roles they play.

Read - Part II: One Size Does Not Fit All: Customizing Self-Service Analytics For Business Users