Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Self-service BI isn't self-enabled

Kimberly Nevala, director on the SAS Best Practices team | Aug. 21, 2013
For many companies, the self-service business intelligence (BI) experience is similar to the early days of agile development.

For many companies, the self-service business intelligence (BI) experience is similar to the early days of agile development. Back then, companies flocked to agile methodologies, drawn by the promise of faster delivery cycles, improved business-IT relations and expedited delivery. 

The method was particularly attractive to those struggling with business user engagement and unmet delivery expectations. What many failed to recognize was this: agile required greater user engagement, more accountability, disciplined decision-making and an ongoing commitment. Not to mention faster, more responsive development. Delivery cycles were shorter and, when managed well, arguably more on point. But success didn't come free.

In the same vein, BI teams struggling under an avalanche of unmet BI needs are embracing the self-service' mantra. Those expecting a quick fix quickly learn the error of their ways.

Self-service aims to make the users' BI experiences more robust and timely. But self-service doesn't equal self-enabled. BI teams are not off the hook. They have a critical role to play in enabling a functional self-service ecosystem. A well-considered strategy must address the following truisms: 

* One Size Does Not Fit All -Just as your customer base consists of different segments with discrete needs, BI users are a diverse bunch. For executives, self-service may mean access to corporate scorecards on an iPad and (or) the ability to "Ask Jeeves" on demand. For line of site business users the best BI is unseen BI. Specifically, reports and alerts that are integrated into not distinct from - operational applications and workflows. Knowledge workers including the emerging data scientist' require access to robust discovery tools and broad data sets. Therefore, enabling self-service requires clear stratification of user types and delivery of a range of capabilities and solutions.

* Different User, Different SLA - Each user segment comes complete with their own expectations and support requirements. This in turn, drives the need for differentiated service-level agreements and engagement models. Not to mention discrete intake and prioritization mechanisms for addressing new requests and resolving issues.  

* A Blank Slate is A Bad Slate - As a general rule, we're all better critics than authors. Self-service works best when users are given a good foundation upon which to innovate. Start with an intuitive, robust dashboard or interactive report or sandbox environment which addresses the most common metrics, report dimensions, analysis and/or data. Then allow users to modify and customize to create variants specific to their needs. 

* Enablement Trumps Development Historically, the unit of BI delivery was a report, dashboard, or perhaps a cube for analytics. Self-service shifts this paradigm from the delivery of a BI "widget" to delivery of data and toolsets to access and exploit that data. Frequently overlooked, but critical to making the turn, is user education. This is far from a once-and-done endeavor. Training and enablement is an ongoing task utilizing multiple delivery paradigms. Formal classroom training. Lunch-and-learns. On-demand user helplines. Enabling self-service is a full-time job. BI teams are not off the hook: the unit of delivery has just changed. Team skillsets and roles may need to change accordingly.


1  2  Next Page 

Sign up for Computerworld eNewsletters.