BI project fail cliff

Common Reasons BI Projects Fail

In our efforts to help organizations develop successful business intelligence and analytics strategies, we have identified some of the most common reasons BI projects fail. Regardless of the size and scale of your organization, these common BI project issues could cause your analytics solution to fall short of its full potential. Like the picture above, avoid these common pitfalls to prevent your BI project from falling off the BI cliff!

1. The team is too technically focused or too business-focused.

A successful business intelligence project requires the right balance of technical and business expertise. An analytics solution that is solely developed by technical staff tends to fail to capture the necessary business processes and expertise required to design a sustainable solution. Similarly, a BI project managed and solely developed by business staff usually fails to understand the necessary technical aspects required for a successful deployment.

A key to success is to have a project team that is comprised of business staff along with technical team members. Cooperation between the business users and IT is absolutely crucial.

2. Too much focus is given to the tools and not the data.

There are many great business intelligence and data visualization tools out there – Qlikview, Tableau, and Cognos are just some of the more popular ones. But organizations just starting to develop their business intelligence solutions tend to focus too much on the BI tool. They tend to forget that a BI tool is only as good as the data that lies beneath it. You can build a car with the body of a Ferrari, but it is useless without an engine. The BI tool is the body, while the data warehouse is the engine.

Successful BI projects focus on developing a quality data model first and the BI tool after.

3. The “I want it all” Problem

Another common issue we find with organizations with a less mature business intelligence culture is that they want every possible piece of data right away. Sure, a successful analytics solution has a lot of data available, but the data is handpicked for a specific reason. If we just try to place all data into a data warehouse or other BI solution all at once, the solution tends to fall apart. Data quality issues often arise and unnecessary data model complexity is introduced.

Remember this important point: a data warehouse business intelligence solution is designed for reporting and analysis. Not every data point is relevant (or at least of high priority) when deploying your first BI solution. Your BI project team should take an iterative approach to developing the BI solution.

4. Data Definition Consistency

An often-overlooked yet critical aspect of a successful business intelligence implementation is clearly and consistently defined data definitions. Different areas of a credit union might have separate definitions for the term “member”. Similarly, banks might have separate definitions for the term “customer”. Depending on a hospital department’s function, they might use a slightly different definition for the term “patient” or “visit”. Without this term consistency, it is difficult to develop a BI solution where pulling a report for the number of members in a credit union could yield five different answers depending on who is pulling the report.

Definitions for key business terms must be defined up front and agreed upon by all project stakeholders. (P.S. this one might end up being harder than you think!)

5. Inflexible Data Models

This is where data warehouse vendor, consultant, or internal development team selection is critical for long-term success or failure. A business intelligence solution needs to be flexible. The data needs of the business today might be very different from the needs of tomorrow. Your development team (either internal or external) should be able to anticipate some of these changes and develop a data model that allows for this flexibility.

The last thing you want is for your BI project to be successful for 18 months and then be deemed useless. Make sure that you build you data model to be flexible as the data needs of your business change.

6. Data Quality and Data Integrity Issues

Another commonly overlooked component of a BI project relates to data quality issues. Anything from bad phone numbers, zip codes, city name spellings or a variety of other data issues can throw a wrench in your reporting. They seem trivial initially, but, if we were to trying to map our customer base using geographic information with misspellings, our information would prove incorrect. As the ETL is developed, ensure that critical details are accurate. Do the number of patients reflected in the data warehouse match the source system? Do the share account balances in the warehouse match the core?

Quick data quality and data integrity checks can save you weeks of headaches down the road. Deal with these issues early before small issues become big problems for your BI solution.

7. Talent Gaps

Do you get the most out of Excel if you don’t have staff that knows how to perform even basic functions in the software? Probably not. The same goes for your BI solution. If no one in-house knows how to write SQL queries, then a data warehouse is limited in the value that it can provide. Some of you will argue that a good BI tool can make up for this deficiency – a point I agree with to some extent, but BI tools can only make up for a fraction of the skill deficiency.

Be cognizant of the skills available internally. Do you have staff capable of developing SSRS reports or reports from another visualization tool? What about SQL – the language used to interact with databases? Answer these questions early on so that a proper plan can be designed to train current staff or hire new staff with the adequate skills.

8. “Only I Can Own the Data”

BI solutions are designed to bring data transparency to the organization. The ability to analyze data across business functions or reporting lines is critical to the success of a BI implementation. Do not fall into the trap whereby a business intelligence or data department holds data hostage.

Allow business users to access and receive information quickly and efficiently if you want the BI solution to be adopted. Hold it too close to your chest, and I can almost guarantee that user adoption will be minimal and any short-term success will be unsustainable.

9. Training, training, and more training

Would you deploy new critical business software without training your staff? Absolutely not. The same goes for deploying your business intelligence solution. Educating and training business users on how and where to access data, the capabilities and limitations of the system, key business definitions, and all other components of the solution are key for adoption.

If there is pushback, then use that pushback as a constructive criticism tool. Create focus groups for users who are not adopting the BI solution. Use these focus groups to figure out how to improve the capabilities and deployment of the system. Remember, your business intelligence solution is going to constantly be changing. Getting feedback from those using (or not using) the tools you give them is the best way to improve.

Like a retail company would never ignore customer feedback, your business intelligence department should never ignore the feedback – both positive and negative – from the consumers of data within the organization.


A BI project is not a short-term project. It is broken into a few phases: pre-deployment, development, and post-deployment. The issues above are some of the common challenges faced by organizations in each of these BI project phases.

Be aware of some of these issues that we brought up, and address them early on to ensure the most success for your BI project.

Posted in DW/BI, Uncategorized.

2 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *