5 Steps To a Successful BI Implementation

Published by

Sucess FailureImplementing a business intelligence (BI) solution can be a game changer for your organization by providing integrated insight into data from all corners of the business. Despite all of its promises, though, an enterprise BI implementation is more often than not met with mixed results (and failures!). This can include horrendous delays, huge budget overruns, data problems, and disgruntled end users.

Some “Gotchas” of Enterprise BI

When BI projects fail, the story often plays out something like this: The development team meets with all stakeholders first to gather requirements. Then they go off to try to accomplish everything and roll-out an entire deliverable all at once. While this approach may sound ideal, it seldom works for many reasons:

  • Delays typically occur because the different reporting areas take forever to agree on what the business rules are for their desired metrics.
  • The data sources or business rules that you assume you’ll need may be incorrect.
  • Numerous data validation issues may exist.
  • The way reports look or dashboards function may be different than desired.
  • Requirements may be misinterpreted and the deliverables different than expected.
  • Delays can happen because everyone’s on different schedules.
  • In the end, it isn’t what your executives had in mind.

Unfortunately, all of this leads to your stakeholders being dissatisfied; business users distrusting the solution; time, money, and resources wasted; and you’re no closer to the BI solution that you needed in the first place.

A 5-Stage, Phased Approach Works Best

An enterprise-level BI implementation works best when it’s done in a staged or phased manner. When you implement BI using a phased approach, there are 5 things you should do to help avoid any major pitfalls along the way:

Gather your analysis and reporting requirements up front: This is a must! It’s a best practice to involve all stakeholders from the executive team to your business users from the get-go. Having all the necessary requirements from them up front saves a lot of extra rework down the road.

Break down the requirements into key business areas and phases: This makes for much more manageable chunks of work, which means it’s easier to stay on point. Plus, project members feel more productive knowing work is getting done. This is opposed to having just one large deliverable. With larger BI implementation projects, when you encounter roadblocks, the delivery schedule can be pushed out innumerable times and you have nothing to show for all the work that has gone into them. Breaking the project into multiple phases allows for a more iterative development process in which interaction with the project stakeholders is frequent and consistent. This not only prompts more feedback that can help you guide and correct your development course, it keeps you from trying to do the whole project at once and then finding out in the end that you went in the wrong direction. Smaller iterations with lots of interaction allow you to gently adjust course when road blocks occur.

Select the priority of phases: Scheduling your phases very much depends on the dynamics and internal politics of your organization. But when figuring out the schedules and priorities, consider the following:

  • Which areas of your business stand to benefit most from a particular phase?
  • Which phases will impact your organization the most?
  • Are certain phases more complex or apt to have additional challenges?
  • Which departments are likely to adopt a new analysis and reporting tool more easily?
  • Which business areas have the time and flexibility to work on a particular phase?
  • Which areas of your business have clearly defined metrics that can successfully implemented right away?

Continually validate data: An iterative approach to enterprise BI is also significant when it comes to data validation. In the world of BI, data integrity is crucial. You can easily lose the faith of your end users forever if your data is determined inaccurate during the initial roll-out. Constantly getting eyes in front of the data allows for more opportunities to catch errors and it makes the final validation/QA stage easier on the validation team.

Gradually roll it out to your departments: Rather than trying to force the whole company to make a complete paradigm shift at once, rolling out the BI application to individual departments and/or business areas allows you to focus your resources a little at a time to make sure the roll out is smooth and successful. Positive word of mouth about the successful implementation and usefulness of the tool can then get other departments excited about their rollouts and more engaged when their turns come.

When it’s all said and done, keep in mind that BI is not a one-time project. It’s a living, breathing animal that requires constant care and feeding. As your business grows and changes, you’ll need to fine-tune your BI system to accommodate those new changes. And having an iterative deployment approach in the wings to support it will make it all the more bearable (and successful) for you!

Categorized in:

This post was written by Pat Hennel