The term self-service BI (SSBI) is common, but it’s also commonly misunderstood. The first few things I ask clients who want to discuss a SSBI project are “What do you mean by SSBI? Who do you expect to use it and what are their analytical and data skills? What data will it access? What are the source systems?” Most importantly, what are the business expectations?

Certainly SSBI fills several important needs, in particular helping business people access data without having to wait in a queue for IT support.  In 2016 Gartner made a major shift in its Magic Quadrant for BI and Analytics Platforms to include only data discovery tools, highlighting an industry shift towards agile and accessible analytics.

But the initial success of a new SSBI tool is often followed up by frustration. Let’s take a look at how that can happen.

A company gets a top-rated data discovery tool. So far so good, right?  It’s a really good tool, no question about that.  Let’s suppose the first rollout of the tool is with a limited set of data, like just Saleforce.com

The business power users start using it.  These are the people who know the business group’s data inside and out, up and down. They know the schema – tables, the columns and all the relationships.  They know what data is available and its quality. They are the experts who built the sophisticated spreadsheets that other business people have historically relied on for information (often referred to as data shadow systems).

These people are really good with the tool. They put it through its paces. They know how to use the tool because they understand the data so well.  The tool is making it easier for them to do all the work they were doing before with spreadsheets.

After the successful rollout, a couple of things tend to happen:

• One, other business people start using the tool. These are the people who are experts in their business area, but not experts in data analysis. They need things like pre-built visualizations and reports that they can pick and choose. They would like to be able to filter the data. They are focused on their area of business; they can’t spend time worry about where data is coming from and what state it’s in. They feel self-reliant if they have the ability to filter the data they use, drill into details, change the visualizations and export it into spreadsheets.
• And two, the tool gets kicked into high gear. It now accesses data from more sources, which means there are a lot more data relationships and a lot more complexity. The whole ball-of-wax we call data integration comes into play, as it becomes apparent that not all the data is adhering to the “5 C’s” (clean, consistent, conformed, current and comprehensive).

The non-power users don’t understand data integration, nor should they. And it’s often more than the power users are accustomed to if they’re not BI experts. (Data preparation tools have emerged as the non-IT data integration-lite tools to fill in gaps, but that’s another discussion.)

The self-service solution, which started out so promising, is no longer delivering the results everyone needs and it isn’t saving anyone time.  None of this is unusual or surprising – it happens in companies all the time.

The key, before implementing SSBI, is to think about these questions:

• What do you mean by SSBI? – Are you thinking of pre-built reports and dashboards or providing the ability to actually build reports?
• Who do you expect to use it? – Your data experts or your average business person?
• What data will it access and where? – A single source of data or multiple sources and types that will need to be integrated?
• What are the business expectations? – That one tool will meet everyone’s needs? That you will no longer need the IT group?

SSBI is a great tool in the BI arsenal, but you need to know what you’re dealing with. When the vendor tells you the tool can solve a host a problems, understand that it’s probably an exaggeration unless your needs are very simple and static.  Set the right expectations and make sure you understand all the nuances.