Improving BI project quality with a Scrum framework

  • Posted on February 17, 2015

improving BI projects with ScrumThe following blog post was written by Avanade alum Tim de Lange.

Companies across all industries struggle to harness the data explosion produced by the pervasive use of digital platforms. Companies are reliant on successful big data or business intelligence projects to harness the value of all this data. As one Gartner analyst said in the firm’s 2015 predictions, “the value is in the answers, not the data." Unfortunately, all too often, business intelligence projects fail to get results.

Quality assurance on business intelligence (BI) projects is not something that is always taken into account when getting started on the development of a new data and analytics solution. But, quality assurance is exactly what’s needed to ensure the project yields results. At Avanade, this is why we are increasingly using a Scrum framework for BI projects where it makes sense. With a Scrum framework, a key foundation is assuring that the quality of the project can be safeguarded.

Mitigating common quality risks

There are many factors that can impact the quality of a project. The most well-known factors are the availability of time, money and the size of the scope. Within a project that uses the Scrum framework, the availability of time and money are no longer an issue. The amount of time is a fixed period as Scrum dictates a fixed duration for each stage that is determined before it starts. And, because the team is assembled before the iteration starts and the duration is set, the amount of money that is spent on the team is also fixed.

Avoiding scope creep

To mitigate the risk of scope creep that could impact the costs, time and quality of the project, the Scrum framework makes use of the product backlog where the product owner can place his requirements. After the product owner has written down all of the requirements – and sorted them by priority, the development team will go through the requirements and create a proper estimate on the required time needed. From there, the development team will fill in as many product backlog items that will fit during each iteration.

The definition of done

The Scrum framework also provides the product owner with valuable tools to ensure quality too – with perhaps one of the most important ones being the Definition of Done. The Definition of Done gives the product owner the power to determine what the product of each iteration, or sprint, must adhere to before it is allowed to be called finished. For example, the product owner may state in the Definition of Done that the business intelligence report must have integration and user acceptance tested before it is delivered. This ensures early quality testing before the product is a final deliverable.

The Scrum master

Next to the development team and the product owner, there is one more role within the Scrum framework that ensures quality control. The Scrum master’s sole responsibility is to ensure the quality in which the Scrum framework is incorporated in the project. Besides monitoring the execution of the Scrum process, the Scrum master is also responsible for resolving any issues or impediments the development team. By doing this, the Scrum master helps the development team keep its focus on the quality of the product rather than processes.

While there is no silver bullet to solve the big data challenge, a Scrum framework can be an effective tool in ensuring that business intelligence projects are on track, within budget and will yield results. As enterprises look to transform their businesses in today’s digital landscape, Scrum can be a powerful tool in the agile path to digital.

Nisha Patel

We have been trying to implement the Scrum for one of our BI project however its getting tough to deal with practical issues of data integration required at each environment which forces us to club all user stories before the testing can begin. I know this is essentially like following waterfall within Scrum :(. While I know the answer is fully automated build, deploy and integration activities I am really looking forward to some case studies of implementing Scrum in BI project where the delivery to project is every Sprint end (we have a 2 week sprint now).

February 17, 2015

Tim de Lange

Dear Nisha,

My apologies for the delayed response to your message on my blog post.
I think the best advice I can give you on this matter is to stop looking at the complete data integration packages as deliverables. All different parts of your package can be a deliverable and can therefore be tested according to your definition of done.
Within the Avanade Data and Analytics service line in The Netherlands we have quite a lot of experience with delivering data integration solutions by using agile frameworks like Scrum.
Feel free to reach out to me directly and I will try to get you into contact with some of our Scrum masters to give you some more practical advice.
Hope this helps.


February 26, 2015

Avanade Insights Newsletter

Stay up to date with our latest news.

Next steps

Talk to us about how we can bring the power of digital innovation to your business.

Modal window
Share this page