6 BI Pitfalls and How to Avoid Them

  • July 10 2009, 10:34am EDT

In today’s organizations, one of upper management’s main goals is to maintain a steady process that provides decision-makers with access to critical business information aligned with corporate objectives. Success in any organization hinges on proper access to this information – not as an annual, quarterly, monthly or even weekly event; but as an ongoing, consistent and standard process that allows information to be readily available to decision-makers on demand. Dashboards and scorecards provide organizations with the ability to monitor performance in relation to goals, obtain status information and identify trends using straightforward metrics or key performance indicators. Business intelligence the skills, technologies, applications and practices that are used to help an organization acquire a more thorough understanding of its commercial context – can provide a historical and a current and predictive view of business operations. When BI implementations fail to convey or inform a decision-maker because the information presented is unclear, it leads to additional time spent getting insight into performance – or even to an incorrect decision made because of an inaccurate interpretation of data.
Common pitfalls can be avoided when designing and deploying a BI solution; likewise, key concepts should be used within a BI user interface to help make a user’s experience more productive. By avoiding these pitfalls and following the recommended design principles, an organization can design a BI solution that will deliver a visual environment abounding with information. This information should be able to be easily interpreted while also providing KPIs, status checks and trends relating to organizational objectives and strategies. A well-designed BI interface enables management to interpret considerable amounts of information in a short time to make business decisions promptly and with confidence. 

Pitfall #1: Too Much BI Information Presented

When BI information presented to users exceeds a single page, it impact how they read and interpret data. When designing BI screens, you must consider that most users can only retain a certain amount of information over a short period of time. Designing BI screens that take up a single page will provide users with the right amount of information to be consumed. In addition, BI screen layout should take into account that people absorb “clustered” information the best. For example, suppose there are two indicators that together provide users with the information needed to make a decision. If users have to scroll the view to see both measurements, they may not be able to link the two together and obtain the insight required to make the correct decision. In addition, users may perceive that the information at the bottom of the screen is less important than the information being displayed at the top.
How to avoid this pitfall:

  • Design your BI layout to fit to a single screen.
  • Break the data up and organize/link it over multiple screens.
  • Group together data that will provide users with the information needed.

Pitfall #2: Ineffective Use of Measurements

Many BI solutions use the measurements that are available based on underlying data. In many cases, these measurements may be irrelevant to users or not what they were expecting. To users, then, these measurements are ineffective and inefficient. For example, most BI solutions will provide a measurement that is comprised of planned and actual values. In most cases, this is presented to users as a line or bar chart to show the values over time. If users are trying to obtain the delta between the two measurements, they will need additional time to determine the difference between planned and actual. It would be best to provide the number associated with the delta instead of making them figure it out. 
How to avoid this pitfall:

  • Design measurements that give users what they need. Don’t force them to do computations the system can handle.
  • Display measurements that are effective and efficient for the end user.

Figure 1

Management is really interested in the ratio of marketing spend to revenue generated, not in the comparison of revenue dollars to marketing cost.

Pitfall #3: Too Much Detail

Avoid providing users with too much information and detail in a single view. In many cases, BI designers want to provide as much information as possible; however, too much data presented at a single time is confusing for users and inefficient. When faced with information overload, users will spend additional time and effort to filter out the unwanted information, while trying to process what is actually relevant to them. Think about how data will be presented to users and the time that could be saved by presenting the information in a more efficient manner. For example, $8.48M is easier to comprehend compared to $8,478,934.99.
How to avoid this pitfall:

  • Keep data at the appropriate level to provide information that can be quickly grasped.
  • Represent data with brief, easily recognizable labels (for example, Q1 09 instead of First Quarter, 2009).
  • Ensure that users can quickly grasp relevant information without having to “filter” it.
  • Limit precision of information (for example, π instead of 3.14159).

Pitfall #4: Incomplete Presentation of Data

Avoid presenting users with indicators that have no context – otherwise, they will perceive the indicators as another measurement and not grasp what they really represent. For example, suppose users are presented with a single measurement: year-to-date gross margin. This bit of information may not provide users with the insight to make a correct decision with confidence. It would be more helpful to include both the year-to-date planned gross margin and the year-to-date gross margin for the previous year. Then users have the information they need to compare current objectives and previous performance to make a more informed decision.
How to avoid this pitfall:

  • Provide the context for comparison that will give users insight into the indicators.
  • Draw more attention to the principal indicator. Do not give equal attention to the context surrounding it.

Figure 2: These gauges reflect measurement, butprovide no context to show if evaluations got better or worse from last period

Pitfall #5: Using Visualizations That Impede User Perceptions

BI solutions provide a wide range of tools enabling visualization of the data. Many visual displays can cause users to be confused about the data and what it represents. Some commonly confusing visual presentations include using charts that don’t effectively represent the data, such as a pie chart to show quantitative data (a bar or line chart would be more effective); colors and contrast that inappropriately draw attention to a measurement that may not be that important; bold type, underlines and borders that divert attention from the data; unnecessary decorations or ornamentation in the BI interface; and too many different visual display types (line, bar and pie charts; KPIs; etc.) in a single view.
How to avoid this pitfall:

  • Use color and font formats to draw attention to the relevant data.
  • Use bar or line charts to show quantitative data.
  • Limit use of unnecessary decoration or ornamentation.
  • Use the same visual displays to present different measurements and provide a consistent method of data interpretation.

Pitfall #6: Inappropriate Location of Data

BI screens should be designed to draw users’ attention to the most crucial and significant data that will give them the insights they need. The BI screen should be planned and developed similar to the way an engineer plans the roads for a real estate development project. If the roads are improperly designed, the flow of traffic is disrupted. If data within a BI screen is not properly positioned, the flow and interpretation of the data will be disrupted.
How to avoid this pitfall:

  • Place the most important information in the upper-left corner of the screen to reflect the reading pattern of most users.
  • Place the least important information in the lower-right corner of the screen.
  • Keep the flow of information on the screen in a logical course.
  • Group information together to maintain relevancy.

When designing a BI screen layout, there are several critical factors to consider. The way that BI screen designers present information can directly influence the way that users ultimately interpret that data. The amount of real estate consumed by the information and the arrangement of that information will have significant impact on how it is perceived by users. Using the correct mixture of visual displays to draw users’ attention to critical data will reduce the time needed to review and consume the information. Finally, providing users with the proper amount of data with appropriate measurements will ensure that they have the right information to make the best decisions for the organization.
The concepts and best practices in this article are based on approaches from the book Information Dashboard Design by Stephen Few. This article also combines my years of experience in developing executive information systems and decision support systems for Federal, state, and commercial clients, including those in the automotive and telecommunications industries.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access