Agile Analytics: What It Is, and 10 Best Practices for Using It

by Chris Preimesberger

Data warehousing and business intelligence systems will live in a diversity of environments, not just in the IT department. It’s important to treat business owners, technical experts, project managers and the many communities of users across the organization as members of the team by allowing them to offer input and test working features as they’re developed.

Because most stakeholders aren’t well-versed in data warehousing and business intelligence, they don’t know what’s reasonable for them to ask or expect, and often they change their minds as they see the system put into action. Investing time in education both up-front and throughout development will help clarify needs and goals, keeping the developed product useful and relevant.

In traditional development models, developers could work for months on a feature, only to find it no longer applicable to a changing business environment. In Agile analytics, each iteration should deliver a working feature to be tested by stakeholders and adapted in further iterations to better suit the organization’s needs.

With so many stakeholders on board, it is crucial to test data warehousing/business intelligence systems frequently throughout the development process. Integrate continuously and test systems in pre-production or demo environments at various benchmarks throughout the project so there are no surprises at the end.

The core purpose of big data is to find key insights upon which an organization can pivot. In this way, big data by definition demands agility. Listen to what users, tests and business conditions are telling you, and work change into subsequent iterations.

The greatest manpower should be saved for developing new features and collaborating across organizational and development teams. As such, it’s important to automate as many regular processes as possible, from testing to administrative tasks so that developers can focus intensely on an iteration’s set goals.

Hire talented, motivated individuals who can set their own goals for each iteration and function as effective self-managers. Then, trust them to do the job at hand, self-monitoring and adapting as they go.

While Agile analytics has many guidelines, it is a style, not a process. More traditional tactics aren’t antithetical to Agile if they’re effective in achieving iterative goals. Choose the tactics that work best for each project and team rather than adhering to static rules.

Agile systems development requires just as much discipline and rigor as the traditional waterfall method in order to stay on track. However, rigor should be applied not to adhering to rigid systems and static goals, but to constantly re-evaluating the effectiveness of the methods and styles at hand.

Keep up-to-date with the best data warehousing and business intelligence practices and implement them fluidly into each iterative phase. This will substantially increase the development team’s agility and keep the organization ahead of its competitors.