News Stay informed about the latest enterprise technology news and product updates.

Categorizing business intelligence users

Part 1 of this article introduced the Baseline Pyramid, which offers a BI taxonomy according to the type of usage and the maturity of the data. In Part 2, we spin the pyramid in order to relate this evolution to classes of knowledge workers.

This article originally appeared on the BeyeNETWORK.

At the phone company mentioned in Part 1, the “eighty-twenty” rule applied: twenty percent of the business community would be doing eighty percent of work on the data warehouse. It made sense to target those users first.

We spun the pyramid around to connect the different analysis categories with different user communities. This helped the phone company and subsequent clients see that different user groups had different priorities in terms of look and feel, business requirements, and depth of usage.

 

Figure 1

In the case of the phone company, more than half of all business intelligence (BI) users would be casual users, who relied on business information circulated on a regular basis via standard reports. These reports tend to be consistent and repeatable. Business users prefer dashboards and scorecards that reflect their organizations’ key performance indicators (KPIs). They see information and take action.

Investigators are more sophisticated, both in terms of their prowess with BI tools as well as with the data. They still need structured data, but they use data in more of an ad hoc way, usually to investigate an anomaly or problem. They are more comfortable with custom queries and the slice-and-dice functionality that accompanies their OLAP tools. Investigators need an answer to the question, Why?

Explorers aren’t sure what they’re going to find, but they know where to look. Explorers try to understand and resolve the details associated with a pattern or business problem. When well-known queries don’t generate results, explorers leverage non-SQL based BI solutions such as advanced statistical or numerical analysis. They aren’t bound by the traditional data structures, but can reorganize the data – and even pull in other data sources – in order to answer their questions. They may choose to aggregate or re-categorize data in order to validate suspicions and mild hypotheses.

One explorer we work with shuns his company’s standard 18-25 year old “target” demographic, segmenting his promotion audience using other attributes, such as purchase frequency.

Not so with the inventors, who are much better at using raw materials to shape new discoveries. Inventors are ready for non-traditional findings. They approach the problem set devoid of any hypothesis. Inventors require detailed data in order to do their work, and are also experts at using advanced statistical and numerical algorithms to discover unknown patterns in the data.

As companies move up the pyramid and their analysis capabilities mature, the data itself not only evolves but also becomes more critical to accurate analytics. Gabriel Tolichoa, Director of Pricing Modeling at Canada Post, emphasizes the increasing importance of enterprise data for his modeling and segmentation work. “If you want business value, you need quick results,” he says. “We let our businesspeople use what they’re comfortable with. As long as they use the right data, the tool itself is less important. It's all about having good, reliable data on demand.”

Many BI best practice companies mix and match their toolsets to leverage the robust data contained in their data warehouses and marts. But these companies always stay mindful of who their users are.

Classifying End Users

In business intelligence, “ease of use” is a relative term. One user’s monthly report is another person’s trend analysis project. More likely, individual users have varying needs for data access and presentation. Understanding and ultimately classifying these disparate needs can make training, business requirements definition, data refinement and enrichment, and vendor technology selection much easier.

Recently we worked with a state government agency responsible for child and family aid. The agency supported diverse data silos across its different departments, and each had its own version of constituent address. But it had recently deployed a geographic information system (GIS) to leverage newly integrated data so that the hundreds of case workers across the state could share information with one another about their cases. With the new system, agency workers could now relate aid recipients to households, school districts, and agency offices.

As part of the GIS deployment, the state agency identified three categories of end users:

  • “On-demanders,” or those who only logged onto the system on a need-to-know basis. This group also included administrative assistants for executives who didn’t want direct access to the data but needed the information through their office workers.

  • “Daily viewers” – usually case workers – who needed access to the data to support their daily work or respond to inquiries from constituents.

  • “Power users,” who would create, maintain, and analyze data in more advanced ways to detect welfare and food stamp fraud, predict future need, and support more advanced analysis on behalf of their case worker colleagues.

By categorizing its end-user community, the state could determine the types of tools necessary for each group – for instance, an intranet-based GIS for distributing maps and geographic data for traveling case workers – as well as individual training needs of each group. Data access rights were also tailored to the end-user categories to cut down on reactive data authorization by overtaxed DBAs. State employees across all three user categories now have access to visual address data for individual recipients, households, and foster caregivers throughout the state.

An IT organization at a banking client keeps a “living” classification of its end users by constantly monitoring business usage of business intelligence and co-funding BI tool and reporting solutions. The organization classifies its users as shown in Figure 2.

Figure 2

For some companies, user classification drives a best-of-breed approach to business intelligence. For others, it validated their plans to expand BI capabilities with additional tool functionality over time.

Meeting the Users Where They Are

Nowhere is end-user categorization more helpful than when the company is considering the adoption of new toolsets. In the bank’s case, classifying different usage behaviors into user groups allowed the bank to reduce the number of BI vendor licenses, while still supporting users’ specific needs for functionality via a variety of vendor toolsets as shown in Figure 3.

Figure 3

But amid the vendor hype, it’s often the new functionality of existing tools that converts an interested businessperson to an earnest end user.

“Our focus was always on the users,” says David Biggers, Data Warehouse Production Manager at Lawrence Livermore National Laboratory (LLNL), a winner of the TDWI Best Practice Award in the Radical BI category. “We focused on enhancing BI capabilities through defined and intuitive business entities, integrated, business-focused metadata, a feature-rich and powerful ad hoc reporting capability, plus a wealth of user-designed canned reports for those business users who would rather not build their own.”

LLNL has deployed increasing BI functionality through integration with Excel, a tool both familiar and functional to the end users there. Lab users may design custom, full function workbooks around corporate data and then allow the central BI facility to automatically refresh and redistribute an updated version with each reporting cycle. Excel, though never intended as a BI “front end” toolset, is nevertheless a popular choice for BI, with many businesspeople using BI standard toolsets as export layers sitting between the data warehouse and the user’s Excel spreadsheets. The concept of meeting the users where they are may mean allowing them to leverage standard and familiar toolsets against unfamiliar data with the aim of becoming self-sufficient with the latter.

The reality is that, as businesspeople continue to use data for operational and strategic BI work, they move to new categories. As corporate strategies evolve and change, new employees join the company, and users become more data-savvy, requirements stay fluid and the need for more data and continued access simply grows. Ironically, as BI tools become more functional, the data becomes more critical.

Dave Biggers from Lawrence Livermore National Labs agrees. “We continue to evolve from one maturity level to the next,” he says. “I’ve been fortunate to work for some very pragmatic visionaries who have understood the confluence of business requirements, technology advances, and the natural progression of business capabilities. We’ve never articulated one vision because by the time we reach a certain level, our corporate vision has changed. To us, BI maturity simply means asking the question, ‘What’s next?’”

Jill Dyché

Jill is a partner co-founder of Baseline Consulting, a technology and management consulting firm specializing in data integration and business analytics. Jill is the author of three acclaimed business books, the latest of which is Customer Data Integration: Reaching a Single Version of the Truth, co-authored with Evan Levy. Her blog, Inside the Biz, focuses on the business value of IT.

Dig Deeper on Business intelligence strategy

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.

-ADS BY GOOGLE

SearchDataManagement

SearchAWS

SearchContentManagement

SearchOracle

SearchSAP

SearchSQLServer

Close