The most common mistakes made by companies attempting to implement business intelligence (BI) solutions today can be traced directly to a lack of understanding of the analytic end users. Very little energy is spent on understanding "how" the end users will use the BI solution; rather, development cycles and user surveys concentrate on "what" the BI solution should do. The key to a successful business intelligence solution implementation is aligning the solution to the appropriate end-user constituencies. To do this, companies must first create an analytic end-user profile, defining who the users are (e.g., information technology [IT], power users, business users, casual users, extended enterprise users) and whether they are consumers of information or producers of analysis. Once defined, this master profile can be used to associate analytic functionality –­ such as OLAP, data mining, statistical analysis, analytic reporting, visualization, enterprise reporting, scorecarding and metrics –­ to individual groups of users, increasing the likelihood of a successful BI implementation. Lacking an appropriate profile of the different analytic end users ­– including criteria for usage –­ IT will continue to build great solutions that are used by no one.

The BI Process Oversimplified

Companies both large and small are implementing BI solutions to gain a competitive advantage by analyzing the data captured within enterprise applications (ERP, CRM, SCM, finance), data warehouses and data marts. The conventional approach used to develop and implement these BI solutions has two distinct phases:

  1. Development –­ focusing on the required data, application logic, analytic functionality and business rules necessary to the process, and
  2. Deployment –­ focusing on delivering a solution to those users impacted by the solution.

Successful projects have included subphases for requirements definition, application specifications, documentation and functionality assessment (oftentimes predicated on information gleaned from end-user surveys). Unsuccessful projects have excluded one or more of these subphases; however, more fundamentally, they lack an evaluation of usability. Missing from this approach is the question: Who will benefit by using the solution, and how do they want to make use of it? The wrong answer is: The end users will get exactly what they asked for. This is wrong because there is no single group of "end users"; rather, end users is an aggregation of all user groups.
In a meeting with a global financial institution, the CIO frowned upon the very mention of BI implementations, suggesting humorously that business intelligence is an oxymoron. When asked what had happened with a project to deliver analytics to a large group of end users, the executive lamented that none of the end users were actually using the solution –­ even after an exhaustive requirements specification initiative had clearly defined what was needed. When asked how many different end-user categories they were targeting with this solution, the answer was one: end users. No thought was given to the differences between analysts, executives, managers, customers or application developers. Lacking an appropriate profile of the different analytic end users, IT will continue to build great solutions that aren't used by anyone.

Profiling Analytic End Users

There are five discrete analytic end-user constituencies: IT, power users, business users, casual users and extended enterprise users. The percentage that each represents varies slightly from company to company (Global 2000 versus small/medium-sized) and from industry to industry (telecommunications versus manufacturing versus financial services); however, these variations are minimal.


Source: Giga Research, a wholly owned subsidiary of Forrester Research, Inc.
Figure 1: Analytic End-User Percentages

A number of Giga clients have commented that IT does not belong within the definition of analytic end users. Giga chooses to include IT in all analytic end-user classifications based on the greater involvement by IT in all BI projects during the past three years. What does change from company to company is the percentage of total users that each constituency represents. In Global 2000 companies, the power users represent two percent of the total analytic end- user population, whereas in mid-market companies ($500 million), the number of power users will be closer to seven percent. The key to a successful BI solution implementation is targeting the solution to the appropriate end-user constituency.

Within the five different user constituencies of the enterprise, there are two distinct groups of analytic end users: producers and consumers. Most BI tools are targeted at the producers, while most of the users are consumers. This paradox has led to companies purchasing the right tool for the wrong group, over and over again. In total, Giga estimates that 14 percent of analytic end users are producers – those who create analytic reports (SQL-based, complex mathematical and statistical processes), build data cubes (OLAP multidimensional databases) and author enterprise reports –­ while 86 percent of analytic end users are consumers of the information and data being provided by the producers. Ironically, even though only 14 percent of analytic end users are producers, more than 75 percent of the costs associated with BI solutions are for technology that addresses the needs of producers. Companies must first determine which users are producers and which are consumers, and then map the available BI solutions to the appropriate users. The producers are generally not decision-makers; however, they facilitate the decision-making process, whereas virtually all decision-makers fall into the consumer category.


Figure 2: Analytic End-User Spectrum of Usage

Survey Questions for End-User Profiling

The following are some of the questions that IT should ask analytic end users to better qualify which users fall into each category and to clearly differentiate producers from consumers:

  1. Ideally, what percentage of your workday would you expect to spend with this BI tool/solution? (Power users, 30-70 percent; business users, 10-15 percent; casual users, less than 5 percent; extended enterprise users, less than 2 percent).
  2. How would you categorize yourself?
    • I produce analysis/reports for others to use.
    • I consume the analysis and reports that others have produced for me. (IT and power users are primarily producers, while business, casual and extended enterprise are primarily consumers; however, there are small groups of producers within each constituency.)
  3. What would be the most useful form of any report to help you in performing your business tasks?
    • Excel spreadsheet (business users).
    • Word document (casual users).
    • Adobe PDF file (extended enterprise users).
    • HTML/DHTML browser file (business users).
    • OLAP cube or relational database table or view (power users).
    • Dashboard or scorecard in a browser (dashboards for casual users, scorecards for business users).
  4. Are you comfortable working with structured query language (SQL), relational database tables and views, or multidimensional cube building? (Only IT and power users should answer yes.)
  5. How would you characterize your job requirements relative to data and information?
    • I need ad hoc access to all data, and have sufficient time to create my own analytic applications (power users).
    • I need data, but do not have the time to create ad hoc queries, nor the inclination to work directly with creating analytic applications (business users).
    • I need information about customers, products, financials and other corporate issues, but have very little time for anything too complex and really only need reports that are presentable, with me choosing only a few parameters (casual users).
    • I need the reports e-mailed to me in a very presentable format on a regular basis, but do not need to create reports (casual or extended enterprise users).

The following are questions that IT should never ask analytic end users:

  • Are you a power user? (No one wants to admit that they are not a power user, yet most will never be power users –­ so don't ask.)
  • Do you want to build your own reports? Or this variation: Do you want IT to build all of your reports? (Either way, IT will be building some of them, so do not look to empower business, casual or extended enterprise users. They are primarily consumers.)
  • Would you use the tool if it works in your browser? (A power user tool is a power user tool, whether it is on a Windows client or in a browser. Determine first if that kind of functionality is needed, and then determine if the Windows client or browser is the proper medium.)

Remember, end users fall into one of five categories. Most end users (86 percent of all analytic end users)– ­ and virtually all decision-makers ­– are consumers of information. Business users will not evolve into power users. It is far more likely that power users will evolve into business users and that business users will evolve into casual users.

Associating the Users and Solutions: A BI Deployment Map

The final step in creating an end-user profile is to associate different BI solution types (OLAP, analytic reporting, enterprise reporting, dashboards, analytic applications, etc.) with the appropriate end-user constituencies.

Before a BI solution is designed, IT must understand which functionality each user needs, rather than which product/vendor each department wants. This will help IT to efficiently rationalize redundant technologies (cost savings) while delivering highly targeted business solutions to more users (increase value).

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