Top reasons your Business Intelligence (BI) project will fail.
Andrew Bourne
Global Strategic Technology Executive | Leading high performing technical teams and innovative SaaS software | Payments | Analytics
Gartner estimates that the majority of Business Intelligence projects fail to reach the intended result and return on investment (ROI). Experience indicates several important reasons for failure. Each issue deserves it's own discussion, I plan to expand on each the future, and what to do about them:
1) Data models are complex. Systems are complex. They become more complex as requirements demand blending data to find correlations, produce reports, design analytics, visualize data, etc. Grasping the meaning of data in your organization requires business context, understanding of the source system, and how the users interact with the systems. This work is not accomplished by one person over a couple of weeks.
2) Dirty data. Data repositories are completely pristine and perfect. That is until the first human enters data. Every data project will have to deal with some level of dirty data. Dirty data can be an annoyance, an impediment, or a show stopper.
3) Tools traditionally available to only large enterprises. Database software, visualization tools, storage, infrastructure, connectivity, those parts of the architecture are expensive, right? Maybe not, more on this later.
4) Poor adoption of user analytics tools across the organization. Picture Grandma in a new Red Ferrari. The data warehouse is complete, users are trained on the new tooling, its all in place for some great analysis. Yet six months go by and nothing. Why not. Often the tools are overwhelming to users. Or the goals of the tools are not clear in the minds of users. So they don't use them.
5) Decision making errors from misinterpretation of information. Does each department view the information through a similar lens? Are there competing agendas? Are goals clearly defined and shared across the organization? Yet those decisions often have implications for the entire organization.
6) BI projects take too long to show positive ROI - Historically, these projects take 18-24 months before meaningful analysis begins. That is too long. The project can lose momentum. The business relevance can be lost. There is a lot of good news here with new methods and tools.
7) Does tooling play a part? Talk a quick look at the last three years of the Gartner magic quadrant for Business Intelligence tools. What trends do you notice? What implications are there for your project?
8) There are other important reasons BI projects fail to realize the full potential including
- Lack of clarity of the goals
- Executive leadership that does not understand/support/etc. business intelligence and the idea of promoting a "information driven culture."
- Others? Fill in the blank____________________
Next, lets explore how to overcome these issue and actually create opportunities for success along the way. Stay tuned...
Co-founder @ Wayzn; Founder @ Outset Career Prep
7yNone of these are the real reason. I've written an article that outlines how the project approach itself is the biggest liability to BI adoption in the enterprise. I hope it is useful. https://www.linkedin.com/pulse/why-analytics-projects-mistake-adam-smithline
marketing :: digital :: technology
9yExcellent summation of BI and the challenges in the marketplace.
Assistant General Manager | Business Process Re-Engineering | Data Protection Officer
9yA great article. many contributing factors from technical and human aspects, which in short, greatly determined by maturity of company, willingness to spend on technology and human capital, etc. all companies do have some form of BI, be it qv, tableau, etc or primitive pen and paper graphs. Effectiveness of approach should be measured within tradition or futuristic companies, on which works best for them.
VP, Cybersecurity Compliance, Governance Risk and Compliance (GRC) Professional
9yGood article. I have run into #2 too often it normally stems from too many application developers not leveraging database normalization best practices resulting in data that is inconsistent making reliable detailed reporting very difficult, time consuming and near impossible at time. #5 has created some interesting conversations in the past for me. It is interesting to see the different answers one can get from different managers when questioned about a specific task or function within an organization. Getting consensus on what does right look like has caused major delays in some projects I have worked on in the past. However, once the common obstacles are passed getting quick deliverable wins can help move a project along as well as gain organizational acceptance.