Vision. Self-service has been an elusive goal for BI professionals for more than a decade. The idea is simple: empower users to build their own reports and dashboards so they get the information they want, when they want it and how they want it displayed. Self-service BI removes IT professionals as intermediaries between business users and the data. This gives business users direct access to the raw material for business insights and liberates swamped IT professionals from a never-ending backlog of report requests. As such, self-service BI offers a win-win situation: Everyone benefits and no one loses. But if that’s true, why do so many companies struggle to make self-service BI work?
Reality. Our research shows that about two-thirds of organizations, or 64%, struggle with self-service BI, giving their self-service BI initiatives a grade of “average” or lower, with 29% rating self-service BI “fair” or “poor.” Only slightly more than one-third, or 36%, gave self-service BI a grade of “good” or “excellent” as shown in Figure 1.
Challenges. Implementing self-service BI is trickier than many BI professionals anticipate. The number one challenge cited by almost three-quarters, or 73%, of BI professionals is counterintuitive: Self-service BI “requires more training than expected.” (See Figure 2.)
How can something be “self-service” if it requires the IT department to train and support users continually? That’s the conundrum of self-service BI. If you want to empower users, you first need to give them the know-how to service themselves, and how to do this varies widely, depending on the skills and experience of the business users. As one survey respondent wrote, “Self-service BI is great for users with analytical experience, but bad for users without an analytical background.” This training not only involves teaching users how to use the tool, but more important, how to interpret the data and which data elements and fields to use for which types of analyses.
Survey respondents also said self-service BI “creates report chaos” (61%), “makes it harder to find the right report” (36%) and the “tools confuse users” (42%). Survey respondents cited other challenges in open-ended comments, many of which reflected the downsides of opening the BI spigot to many users, including poor query performance, lack of adequate access control, an explosion of nonstandard BI tools and increased licensing costs. Other issues were germane mainly to power users, such as lack of analytical flexibility, missing data and runaway queries caused by users trying to download large sets of data to an Excel spreadsheet.
Making Self-Service BI Work. There’s no “one size fits all” in self-service BI. Different types of business users require different types of self-service approaches and tools. But this simple fact eludes most BI professionals and business-side sponsors. The biggest mistake most BI teams make is buy a single self-service BI tool and give everyone in the company access to it. That’s a recipe for disaster.
The reality is that the vast majority of business users find self-service BI tools confusing. As a result, they quickly revert to old habits like asking the IT department to create custom reports for them whenever they need a different view of an existing report. There is a ton of self-service BI shelfware in our industry. The primary culprit is this mismatch between users and tools.
The quickest way to right a sinking self-service BI ship is to recognize that there are two types of users, each with very different information requirements. There are casual users, who use information to do their jobs, such as executives, managers and front-line workers; and there are power users, who are hired to analyze information. They include super users, business analysts, statisticians and data scientists.
Casual users. The most important thing to remember about casual users is that they have very basic information needs. Most simply want to view the output of a report and perhaps click a few times to see more detail. As a general rule of thumb, 80% of the time casual users want basic information interactivity delivered through a canned report or dashboard. But 20% of the time, they want true ad hoc access to information so they can create their own reports and dashboards. This 20% ad hoc usage makes BI professionals think that casual users want to author their own reports. This is categorically not true, and it’s the biggest reason why self-service BI initiatives fail. For BI professionals or people raised on technology, this is hard to fathom.
To best meet the information needs of casual users, BI professionals must adopt a hybrid approach: they must deliver departmental dashboards to meet casual users’ interactive requirements and personalized support to meet their ad hoc authoring requirements. By personalized support, I mean super users who reside in their department and can be called on to create custom ad hoc reports for their colleagues.
Power users. Power users are the exact opposite of casual users when it comes to self-service BI. Rather than restricting the options of power users, BI managers need to liberate them. Power users, by definition, need to explore data and generate answers to a wide variety of questions that cannot be anticipated. Often, power users need data that does not exist in the data warehouse to answer a business question. And for most power users, BI tools don’t provide the flexibility they need to manipulate data to answer complex questions. This poses a big problem for BI managers who want to standardize on a single self-service BI tool for all users: a single BI tool is either too powerful for casual users or not powerful enough for power users. It’s a no-win situation.
Self-service BI can empower users and increase BI adoption, but it is difficult to implement properly because there are many types of users with different information requirements. There is no single tool or approach to self-service BI that works in all situations. To succeed with self-service BI, organizations need to take a nuanced approach that maps users to tasks and tools.