Avoiding Pitfalls in a BI Initiative

By Sherman Kong, Business Intelligence Practice Leader,
Broadstreet Data Solutions

“87% of BI-related initiatives in the United Kingdom alone have failed to live up to stakeholder expectations”

Most organizations have been pitched the hypes and values that Business Intelligence (BI) brings to the table. BI is widely recognized as an effective practice of data collection and Information Access. Knowledge would be modeled and fed to different entities of a company to streamline business analysis and in turn improve the efficiency of its daily operations and strategic planning. Many organizations continue to spend heavily on projects to upgrade the underlying technology that supports their BI solutions, as well as the labour and time it takes to shape and govern their business processes in an attempt to fully maximize this concept and leverage their investment. Yet statistics from a recent National Computing Center survey indicate that 87% of BI related initiatives in the United Kingdom alone have failed to live up to stakeholder expectations*. The end result is that customers and industry experts are realizing that the outcome of any tangible return on investment (ROI) and business benefits that BI and analytic technologies have proposed to offer is often of a hit-and-miss nature. What has gone wrong in the typical implementation of a BI project to derail its success? Let’s examine some common pitfalls and potential countermeasures to avoid such failure for your organization.

* Source: Lawrence, Dave. “Business Intelligence - Results of the Rapid Survey” 2006. The National Computing Centre Ltd.

DO: Have a long-term vision
DON’T: Be obsessed with short-term milestones

“A vision needs to be established first in order to properly measure and track the overall success of the initiative”

Many BI projects stem from one or a set of functional needs related to capability issues that users have experienced with their current information processing tools. Short-term milestones are then established to track and measure the success of defined deliverables in meeting user expectation. This is a common mistake. Customers often get bogged down with the specific details and functionalities of the tool too early on and do not realize that a vision needs to be established first in order to properly measure and track the overall success of the initiative. Those issues are certainly important but they should not kick-start a BI project. The overall objective should not pertain to enabling users with the flexibility of ad-hoc querying or guided analysis through an elaborate dashboard solution, but should instead be directed to a mission statement that compliments the company’s long-term performance strategy. It is only when its business units see the value of the BI tool in mitigating risks, helping to discover revenue opportunities, or solving a similar business question or concern that an organization becomes proactive in the analysis of their collected data, rather than in a reactive manner. Customers will also be better at evaluating the tool from more of a strategic perspective, aligning the technology and even so the evolution roadmap of the vendor’s products to the timeline of their business goals.

DO: Cost saving
DON’T: Cost cutting

“Cutting expenses from a purely cost-reduction standpoint will not guarantee long-term success”

Most companies these days are far too focused on unpredictable fluctuations of the economy and therefore resort to trimming their budget to help minimize overhead, and cut costs. BI initiatives often become a victim in this game and projects are either delayed if not cancelled. But cutting expenses from a purely cost-reduction standpoint will not guarantee long-term success and thus companies will not be equipped with the right tools when the economy is revived. These customers fail to go back to the basics – that is to reintroduce their mission of strengthening their business values. Once organizations realize that there is an opportunity cost associated with not answering the questions of “how to forecast the company’s strength in the years ahead” or “how to maximize the market penetration”, they will understand that a cost-saving approach is more effective. This usually means revaluation and restructuring of project plans to maximize in-house resources for development while utilizing external domain expertise to act as an advisory role or participate in more intricate initial design phase. This would ensure limited knowledge gap from hand-off or transitioning to implementation. Logically it makes sense, customers have already procured the software licenses, and letting them sit on the shelf as they slowly approach expiration or even end-of-life support dates will only lead to more losses.

DO: User acceptance
DON’T: User submission

“Waiting until the project is nearly completed before unveiling to the audience has seldom proven absolute success in securing end user buy-in”

One preventable, yet still commonly overlooked issue in BI projects relates to frontline user resistance to the end product introduced to them. This stems from a mistake made time and again during project planning and execution. An assumption that communication to end users about the end benefits alone is sufficient to convince them to adapt to the new tools or changes to their existing solution once introduced, and hence waiting until the project is nearly completed before unveiling to the audience. This has seldom proven absolute success in securing end user buy-in.

When it does, it is only because they have been submitted to enforcement by upper management. In the case where the end user group was senior or executive staff themselves, they were not involved earlier in the project either because of their lack of time to participate, or they were not asked by the project team as team leaders out of fear that the act of doing so will exhibit as a lack of skills in delivering the expected solution.

However, user involvement throughout the duration of the project is important to prevent it from falling out of scope and allowing more time to make changes based on user feedback. The misconception of user involvement prolonging the project and posing a negative impression needs to be addressed and understood across all levels of communication from the very beginning. Certainly there will be setbacks as revisions take place based on feedback, but validating business rules or look-and-feel of the report interface at the end of a project would better prevent resistance from occurring. It also leads to under-utilization of the tool as the audience resists change. Those reports and dashboards might be deemed useful for a period of time but afterward users go back to the old way of exporting data into spreadsheets for manual manipulation, resulting in a hurdle of disparate spread marts across the organization with each working off of inconsistent data.

DO: Synergy
DON’T: Unbalanced scorecards

“Teams onboard of departmental initiatives should still keep in mind the connection between their own operations to the overall mission or process drivers in the company”

Another common problem arises from the inability of customers to synchronize their business processes and definitions across the organization. Subunits or business sectors within their company may be at a very mature stage in their respective BI curves, but they could not seem to get over the hump of consolidating their individual progress together to form an enterprise-level view to analyze the overall company health. This is usually due to the phenomenon of business rules and/or management approach existing in silos. The number of definitions of a simple key item such as “product” or “supplier” in a sizeable corporation is rarely just one. Similarly, cost allocation models might be different depending on the function of the department. How should one distribute overheads, office space costs? Should it be associated to the country, region, or business unit levels? This poses a bottleneck when trying to create harmony between units in coming up with a single, company-level, sales figure. It is true that not all BI projects are at an enterprise level but teams onboard with departmental initiatives should still keep in mind the connection between their own operations to the overall mission or process drivers in the company. Also, whenever commonalities are spotted, an effort should be made to fix the proliferation of definitions and streamline the flow of analysis from a top-down perspective.

This will easily accommodate for future growth of their BI solution once it reaches the requirement of a 5000-feet bird-eye view. Additionally, it will demonstrate potential to executives in helping them oversee their operations much more efficiently. Furthermore, requesting more time and funding may no longer seem to be an impossible task. This ultimately ties back to the point made earlier about acquiring a long-term vision.

In summary, customers should first recognize the value that a BI solution will bring to them before considering the specific technological details. Emphasis on planning is imperative to the success of a BI implementation, and that does not only imply scoping out the technical design and development but also all the events prior to it: establishing a project management framework in which objectives can be measured and justified, recognizing user involvement as an important component to the solution, and identifying where inconsistencies in business processes can be addressed.