The self-service business intelligence (BI) genie is out of the bottle.
In Forrester's new book, Empowered, Josh Bernoff and Ted Schadler write about how businesses must empower their employees to leverage emerging technologies specifically social tools, mobile, video and cloud computing to serve an empowered customer base. BI is a key technology for HEROes highly empowered and resourceful operatives not only helping them to make sense of the mountains of data that they have to deal with, but actually allowing them to make smarter decisions faster.
But BI applications tend to exhibit an interesting paradoxical pattern: the more you use them, the more new and different requirements keep pouring in. This never-ending snowball effect of new BI requests from business users puts a significant strain on IT resources. Even with the most noble IT efforts, the "build it and they will come" BI paradox will take its toll.
Rather than fighting the never-ending battle of self-provisioned BI applications, BI business users have no choice but to start fulfilling a significant portion of their own BI requirements using BI technologies that can enable BI self-service and empower BI HEROes.
But how does one become an enterprise BI HERO? The traditional BI support model with heavy reliance on IT is just not sustainable in the long term. As a result, more and more organizations turn to tools and technologies that will allow them to reduce dependence on IT and enable BI self-service by business users. Indeed, self-service features of BI applications can turn a "BI requester" into a BI HERO. But when trying to self-serve most or even some of their daily BI needs, these HEROes run into a different challenge for a couple of reasons:
- "User-friendly" can be a misleading term. "User-friendly" for whom? What's intuitive to a BI professional is not necessarily intuitive to, for example, a marketing analyst.
- Integration with desktop office applications is not enough. Many BI vendors will also tell you that simply by the virtue of tight integration with desktop-based office tools, self-service functionality is virtually guaranteed. While important, it's not nearly enough.
- Point-and-click GUI is not a panacea. Its a step in the right direction, but only one of many foundational components necessary for BI self-service.
Let's consider the entire universe not just the "intuitive" GUI of end user self-service BI requirements. What tools and capabilities does one need to become an effective and efficient BI HERO?
To begin, we need to clearly separate the BI self-service needs of the average or casual user from advanced or power BI users. Casual BI users typically need to run and lightly customize canned reports and dashboards, run ad hoc queries, add calculated measures to existing reports, collaborate with other users, and fulfill their BI requirements with little or no training.
What BI tool features do enterprises need to support such self-service requirements? While most of these are commodity features today report and dashboard templates, customizable prompts, sorts, filters, and ranks, as well as report, query, and dashboard-building wizards some others are not as commoditized, and require more explanation:
- Portals are key enablers of BI collaboration and knowledge management. IT organizations very often end up building thousands of canned, standard, production BI reports. Unless you know what's out there and where to find it, these reports may end up just idly sitting there, wasting valuable IT resources. Portals can help address this challenge.
- A semantic layer isolates business users from DBMS complexities. Understanding the complexities of underlying database structures, often with cryptic and meaningless table and column names, confusing aliases, and little to no correlation to the business question, the business problem at hand is definitely not a task a casual BI business user can tackle. If done right, this semantic layer will present data to a report query builder in clear, descriptive, relevant, and easily recognizable business terms.
- ROLAP and prompting for columns can reduce the number of reports and data marts. Prompting for a column that is not prebuilt in a report and creating virtual cubes (ROLAP) that can point anywhere in any of your data sources significantly reduces the number of reports and data marts that need to be maintained and, therefore, reduces reliance on IT and improves BI self-service.
- BI with search UI can address the I-don't-know-what-I-don't-know dilemma. In addition to requiring little to no training, a search UI for BI effectively solves one of the oldest dilemmas in BI: having to know exactly which questions to ask to get a meaningful answer.
- BI with search UI can also support faceted navigation. Faceted navigation allows analysis on data structures that conventional OLAP tools can't easily provide.
All of the BI self-service requirements of casual or average users also apply to advanced users, or "power" analysts. However, this segment of BI HEROes also has its own more complex set of BI self-service requirements. For example, these power users often need to analyze multiple what-if scenarios, explore entity relationships and add hierarchies not supported by the underlying data model, provision new data sources on demand, and provision entire BI applications on demand.
Forrester recommends a three-step approach to ensure that your BI environment, platform, architecture, and applications can indeed support broad self-service requirements:
- Define BI self-service requirements by user roles. In addition to defining self-service requirements according to casual and power users, identify similar requirements for other use types and roles, such as executives, or analysts using advanced analytical tools.
- Ensure that your BI platform supports all of the commodity self-service features. Most of the leading BI vendors covered in the 2010 Forrester Wave evaluations of the BI market support all of the essential features required for self-service but, if you use products from smaller vendors or BI software-as-a-service (SaaS) subscriptions, do check to make sure you get the minimum set of such basic features.
- Supplement commodities with differentiated features. Not even leading BI vendors have all of the features necessary to cross all of the t's and dot all of the i's for the ultimate self-service BI functionality, which is why Forrester often recommends using more than one BI vendor for complementary self-service functionality.
Self service BI should not be viewed or used as a replacement for traditional BI approaches. There will always be requirements for mission critical BI applications that require heavy IT involvement, ensuring data quality
and single version of the truth. Rather, self service BI approaches and technologies should complement traditional IT centric BI. For example, in certain customer facing scenarios, such as sales, marketing, customer support, giving a quick and timely answer to a client, even based on semi-trusted data, is sometimes preferable to giving an answer based on fully trusted, approved report, but a week, or even a day later, when its already too late and irrelevant.