Understanding the Two Types of BI Users
Self-service business intelligence (BI) promises to empower business users to answer their own data-related questions without help from IT, but any individual solution’s ability to deliver on that promise depends on its accessibility to all types of BI users.
In order to evaluate a solution’s usability, it’s important to have a realistic understanding of your user base and its makeup. We’re going to use research published in Eckerson Group’s “A Reference Architecture for Self-Service Analytics” (2016) to describe the typical SaaS user base and identify the types of tools each user group is likely to require. Compare your user base to this model, and be sure to check off the appropriate boxes before committing to an embedded BI platform.
Classes of Power Users
Let’s start with the minority, the power users. According to Eckerson Group, power users “are hired to collect and analyze information on a daily basis” and make up approximately 10% of all “knowledge workers” in an organization. The power user’s specialized knowledge of databases, querying languages, statistical analysis, and modeling distinguishes them from the casual user.
Eckerson identifies at least two types of power users: data analysts and data scientists. The major distinguishing characteristic between the two is that data scientists are equipped with computer science backgrounds whereas data analysts are not. Data analysts, typically hired by department heads, “might create pricing plans, performance metrics, budget estimates, demand plans, or retention models” while data scientists work with raw data to create, for example, predictive algorithms.
Serving the Power User
Self-service BI needs to cede maximum control to power users without granting them access to administrative tools. This is especially important in BI-enabled SaaS applications, as only the SaaS provider should have administrative access, regardless of how many power users its customer companies employ. Key BI features for power users include:
- Custom Data Modeling: The default join path might not be the preferred one, so make sure power users can tweak data models on a report-by-report basis.
- Advanced Sorting: Instead of simply sorting a data field alphanumerically, power users may want to manipulate that field in some way and then sort it. In practice, this might look something like sorting a date field by day of the week.
- Advanced Filtering: Whereas basic users might be satisfied with aggregate filtering (apply filter A, then filter B, then filter C, etc.), power users may need to introduce some more complex grouping logic to accomplish their goals (for example, apply filters A and B, or, if no data is returned, filter C). They may also want to filter a manipulated field as in the sorting example above, removing all Mondays from a date field by applying a weekday function. Lastly, since power users often create reports for casual users, they may need the ability to present those running the report with a limited set of prescribed filtering options.
- Detail Row Access: It might surprise you to learn that not all BI solutions give users access to detail information. This level of granularity is crucial, especially for data scientists.
- Advanced Visualizations: Power users will need more than the basic bar chart to do their jobs. Look for the ability to apply both static and data-driven benchmark lines to visualizations as well as support for multiple y-axes. Check for chart types your users are likely to need, such as geomapping for retail distributors.
- Event Handling: In select cases, power users may need control over the application’s programming itself. For example, they may want to retrieve a report’s SQL query when it hits the database without running to an administrator for those details. A toolbox of event extensions can prove a powerful resource.
Even though power users are most likely to be the ones agitating for self-service BI, it’s equally important to ensure the solution you choose caters to casual users, as they make up the majority of knowledge workers.
Classes of Casual Users
Casual users, which make up a whopping 90% of the work force, “use information to do their jobs” but were not hired for the express purpose of analyzing data. Most casual users have little-to-no experience with database systems; many are executives, managers, and frontline workers.
“The promise of self-service analytics is not to eliminate IT from the equation, but instead foster greater collaboration between business and IT.” – Eckerson
Eckerson subdivides casual users into two additional types of business intelligence users: data consumers and data explorers. Data consumers, as the name suggests, “simply want to consume reports and dashboards created for them.” In this case, “consume” refers to running, exporting, and interacting with dashboard and report output. Interactions might include such superficial changes as adjusting a filter value, drilling into a report, or sorting a data visualization.
“Data explorers,” by contrast, “are data consumers who occasionally want to edit a report or dashboard or create one from scratch without coding.” Both the occasionally part and the without coding part are crucial here. Data explorers are still casual users whose professional focus lies outside the BI space; they have neither the bandwidth nor the training to design highly complex reports. They can, however, create a new dashboard using pre-existing reports, drag-and-drop fields to do cursory data discovery, and select from a list of aggregation types to perform basic calculations.
Serving the Casual User
Although casual users will still need to be trained to use BI effectively, you want the barrier to entry to be as low as possible. As an administrator, you also will want the ability to restrict user access to advanced features, a tactic designed to prevent the casual user from becoming overwhelmed by and/or frustrated with the application, which can lead to low adoption rates. Key BI features to support casual users include:
- Schema Aliasing: The ability to alias data object and field names goes a long way to making datasets easier for casual users to navigate.
- Data Discovery: Casual users value being able to see and explore their data as they begin to manipulate it. A tool that exposes literal values rather than abstractions of those values can be a powerful discovery tool.
- Default Data Modeling: Casual users typically don’t have enough information to create data models from scratch, so it’s important to have a BI solution that will apply context-sensitive default models for them.
- Drag-and-Drop Design: Drag-and-drop UI controls mimic tactile experiences, which help casual users grasp more abstract concepts. Exago BI has a casual user report building feature that, for example, allows the user to group by a data field simply by dragging it. The motion mimics the placing of that field ahead of the detail, helping teach the grouping concept while glossing over the backend mechanics involved.
- Automatic Charting: It’s easier to edit something that already exists than to create that something from nothing, which is why it’s valuable to have a solution that can summarize a report at the click of a button. Users can then make adjustments to the chart as needed, learning from the chart algorithm while getting their work done.
- Intelligent Formulas: Formulas have a syntax all their own, so when casual users need to perform a more sophisticated calculation, it’s helpful to have a formula interface that guides them through the process. It’s also useful to have common functions (e.g., totals, date formats) displayed as buttons or menu items so that casual users can apply them without thinking about syntax at all.
And of course, it goes without saying that none of these UX guidelines function without facilitation from IT. Admins need the ability to dictate which users will have access to which features and make changes to those configuration settings as knowledge workers transition from one user class to another, as sometimes happens.
Adopting a self-service BI solution that can meet all types of BI users where they are dramatically improves your project’s likelihood of success. As Eckerson puts it, “The promise of self-service analytics is not to eliminate IT from the equation, but instead foster greater collaboration between business and IT.” When each user group has access to the right business intelligence tools, it boosts the organization’s overall efficiency.