We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.
Start a Discussion

What's the best way to prevent business intelligence failure?

Vote 0 Votes
BI that works creates real business value, but according to this blog by Herman Heyns, Business Intelligence: avoiding intelligence failure, half of all BI projects fail.  So what's the best way to prevent BI failure?

5 Replies

| Add a Reply
  • user-pic

    Best way to prevent BI Failure is to assume that Users know precisely what BI they need when you ask them for Requirements! Many BI efforts assume this, collect what they think are requirements from various departments and functions and build a BIG BOX, centralized effort. Users then start realizing that they have left out BI that's truly useful for them at any time and ignore the centralized Big Box effort and start Spreadsheet based Skunkworks. The best BI efforts are built iteratively since users need to see what's possible, try out some BI, be allowed to change their minds and allow for unforeseen BI requirements at any time.

  • Think big and start small.
    Deliver in cycles and refine the system along the way.
    Do not over-promise and manage expectations.
    Do not oversell.
    Think, think, think and then think again, wear the users' shoes and even sleep in them.

  • BI projects fail in the requirements phase itself on most occasions. Failure can be averted by capturing detailed user expectations with functional and non-functional requirements and then getting them duly signed off. Properly planned Proof of Concepts and Technologies can also help in averting technical failures.

  • BI projects usually fail because BI is not something that can be "developed." Infrastructure, data discipline, etc. can be set in motion with classic IT project funding and management, but the true benefits of BI (using the term in the broadest sense) develop over time because no one really knows in advance what is needed. For that reason, BI "projects" are forced to start with replacing one or more reporting systems, or merely thrown out with the hope that people will pick it up. This never happens.

    BI tools and rollout ideas usually come from the vendors and/SI's and are therefor heavily weighted toward lots of (expensive) software licenses and lots of (expense) development.

    Another problem is "best practices," which, like the above, are not developed in the interest of the business users or even the clients.

    BI can be rushed into place to fill a gap or solve a problem, but neither of these situations creates the kind of learning and process that lead to long-term BI success.


  • Business Intelligence, like many other IT disciplines, is governed by the classic triumvirate of People, Process & Technology. Having said that, the nuances within these 3 areas makes or breaks a BI program.

    Without getting into details, here are some points that can ensure the success of BI programs:

    1) People - It is important to realize that BI programs requires many different (often disparate) skills to make it successful. Program / Project Managers, Domain Experts, Business Analysts, BI Tool/Stack specialists, Architects, Data Modelers, Usability engineers, Testers etc. need to work synergistically to accomplish BI project goals.

    2) Process - BI programs are always in an evolution mode - it is not completely 'done' at any point in time. So the process & methodologies adopted needs to accommodate the paradigm of concurrent support & development and in delivering bursts of business functionality at regular intervals of time.

    3) Technology - The key question here is: What is the fitment of BI technology & tools to business imperatives? Business imperatives are related to business model, strategy, value creation etc. of an organization and the tools should support the technology manifestation of business goals/objectives.

Add a Reply

Recently Commented On

Monthly Archives