This content is part of the Essential Guide: How to choose manufacturing software for business intelligence
News Stay informed about the latest enterprise technology news and product updates.

Building a business intelligence software implementation checklist

Follow these best practices to ensure success for your BI implementation.

Implementing manufacturing business intelligence software, or BI, is a good way to get a clearer understanding of how a business works and its overall health. Before jumping in, however, companies should review recognized best practices for a BI implementation. As with any new software system, there are a number of traps awaiting the unwary BI implementation team.

The first critical item in a pre-implementation checklist: Ensure that your BI implementation aligns with and supports the corporate strategy. “You have to be clear on the reason you are implementing BI in the first place,” said Howard Dresner, chief research office of Dresner Advisory Services in Nashua, N H., and author of several books on BI. “It can’t just be about giving the executives fancy dashboards. It needs to be tied to something more substantial like gaining true perspective on how the business is doing to help move it forward.”

Once a company has cleared this basic hurdle, the technical work of the BI implementation begins. Prepping data to populate the data warehouse against which queries will be performed is no small job. Every organization has “dirty” data, according to Dresner, meaning duplicate, inconsistent, inaccurate or incomplete data. Manufacturers must improve the quality of their data before proceeding, or the whole endeavor will be for naught. A business can’t make better decisions based on faulty data.

“Most organizations don’t trust their data,” said Dresner. “You have to be sure that the data foundation your system is built on is rock solid. The business people have to be able to rely on that without any doubt or reservation.”

Even if the IT staff has experience with data cleansing, experts recommend that companies budget more time for cleansing the data and feeding it into the BI tool, which is often called ETL -- extract, transform and load. It is likely that companies will need outside help at this stage, as well.

Pace BI software implementation plans

As with other types of software implementations, don’t try to do it all at once, advised William McKnight, president of the McKnight Consulting Group, in Plano, Texas. First work on a significant pain point, such as understanding which orders did not ship on time, and then gradually expand the BI implementation.

“You can bring in more data and reports over the course of time -- that’s easy to do with BI,” said McKnight.

At the same time, McKnight cautions against taking short cuts that will reduce the utility of the BI analytics tool. For example, while populating dashboards or scorecards, make sure users can click to drill down and learn more about a particular trend. “Users need detailed data, not just summary data. Even if you don’t see the need for that at first, you will,” said McKnight.

Before designing a data warehouse, “know your users’ access patterns and model them to deliver high performance. If not, your users won’t trust the data and they won’t use the system,” McKnight suggested. Unlike traditional analytical software, modern BI tools are designed to allow users to serve themselves, he added, so IT managers need to make it as easy as possible for that to happen.

The most important practice when it comes to BI implementation, experts say, is to make absolutely sure the business objectives -- and business users’ needs -- come before the technology. IT staff can get carried away wanting to implement new bells and whistles. Experts recommend that businesses always tie back each feature to how it will help drive business objectives.

Dig Deeper on Manufacturing management

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.