We're in the early stages of evaluating a business intelligence (BI) product. Can you provide some tips for the requirements-gathering phase? How can we help this process run smoothly in our organization and get the necessary input from vendors in order to make the right decision?
Good questions. If you don't know what you want out of a business intelligence (BI) or related tool, then you are not likely to get it. For starters, understand both the quality and layout of your current data. online analytic processing (OLAP) tools can slice and dice data in many ways. Hyperion, Cognos and Microsoft are just some of the vendors that make very powerful data mining tools but, regardless of the product selected, looking inward first is a very good idea.
Think about the data that moves your business: financial, HR, sales and other types. BI and master data management (MDM) tools can analyze these data and manifest hidden trends in your business but, if your data is a mess, then these tools won't magically clean it up. Once you understand your data, you will better understand what you want out of a potential solution. You might want to embark on a data cleanup project before you purchase any tool. Garbage in, garbage out.
Second, consider your current reports. Which reports are essential and which can be retired? BI tools open completely new doors but some people are only interested in seeing the same old cuts of the data. Some organizations simply use BI tools to recreate reports and don't fully utilize the power of their new toys.
Third, take vendor promises with more than a grain of salt. While their tools are probably a major step up from what you currently have (if not a quantum leap forward), it's unlikely that out of the box any one tool can meet each business need without some type of customization of expensive add-on. It's best to honestly reassess each reporting need to determine if it's truly essential.
Fourth, look outward to other organizations in your industry to see what they have done with similar tools. Learn from their mistakes. Go to user-groups or engage in online discussions with them.
Finally, think about your current resource levels. Buying an expensive tool when end-users don't have time to spend with consultants to configure the tool — much less learn how to use it — is a recipe for business intelligence disaster. Make sure that you have people who understand how the tool can be developed because, despite your best efforts, your organization will not initially define all of the business requirements. Moreover, requirements change. You certainly don't want to be beholden to an expensive consulting firm to make the slightest modifications based on changing business requirements.
Dig Deeper on Business intelligence software
Related Q&A from Phil Simon
If you're trying to build a data quality business case, make sure you know common issues caused by poor data quality, with tips and advice from a ... Continue Reading
A custom system design can pose unique challenges. Learn why and and find out the difference between a customized and non-customized system ... Continue Reading
Get a snapshot of the open source data management software vs. on-premise tools debate with advice from a systems expert, including creating policies... Continue Reading