As Boris Evelson, VP and principal analyst at Forrester, explained in his breakout session at TIBCO NOW, everyone is doing agile analytics, just not talking about it—because they’re doing it in Excel. Given that state of affairs, how can BI professionals help their companies become more agile and support mission-critical needs in “the Age of the Customer”?
The Age of the Customer
In an age where an upstart like Uber can improve on the wait and expense of the traditional taxi service, everything has become commoditized. Services and goods are so much the same that only the label differentiates them. How do you compete when customers have ubiquitous and constant access to your and your competitors’ products and services? Through its research, Forrester believes that business agility—the quality that allows an enterprise to embrace market and operational changes as a matter of routine—is necessary for success, and analytics is a key component.
Evelson observes that employees don’t use enterprise BI platforms when they are too cumbersome and don’t provide the right quality, data models, and relationships between entities. Instead, they choose Excel, which may be agile, but not scalable, resulting in a huge lack of agility for BI professionals, businesses, and customers, too.
Agile BI has four components: software development, organizations, processes, and technology. Unless you have some best practices for all four, you are probably not going to be as successful as you need to be.
Agile Software Development
Evelson recommends rapid development of software prototypes, tangible deliverables every couple of weeks followed quickly by user responses that determine the next iterations. Small teams can pull up Spotfire Cloud and collaborate to come up with a fast solution that meets requirements within a few minutes, with no need for documentation.
Although it may be politically incorrect in some organizations, silos are infinitely more agile than shared services, notes Evelson. In a centralized support organization, you share resources and save budgets, but you’re not going to be agile with multiple sign-off levels, steering committees, and differing priorities. He suggests a middle-of-ground approach in which you empower individual contributors—and do not outsource business intelligence because you need close interaction to build a rapid proof of concept of your analytic model. That model should map closely to the business. Customer-facing concerns like Sales and Marketing have more urgent needs than HR and Finance. Make sure your BI environment can respond to both appropriately.
Start with sandboxes that let individuals connect to whatever data they need, analyze using Excel or better, and share easily with as little control and management as possible. Good tools make a sandbox highly self-service using natural language processing, text-to-query, voice-to-query, map overlays, mobile access, and other intuitive features.
A shared BI environment offloads tasks from IT, while the role of BI professionals, much like a competency center, is to monitor what users are doing and set parameters for red flags. If an analysis incorporates several mission-critical data sources involving complex queries on billions of lines of data being shared via SharePoint, that’s a flag. Professionals can jump in and operationalize to make it faster and easier.
Agile BI Technologies
Sandboxes and agile processes supplement, but don’t replace, what BI professionals do for the enterprise. You still need technology to make the environment more agile, and Evelson suggests choosing from among a stable of “cold, warm, and hot BI technologies” to make analyses faster, better, and cheaper.
Schema versus schema-less data models, and large versus small datasets, help determine what type of processing can be done in the data warehouse (cold), in Hadoop (warm), or in memory (hot).