Recently my colleague Javier Guillén wrote an excellent article detailing some of the challenges and best practices relating to implementing self-service business intelligence (BI) strategies on this site. For those considering or who have already embarked on a self-service oriented BI strategy, his insights—based on numerous real-world implementations—are a must-read and should be taken to heart.
Self-service BI is a relatively new idea and one might rightly ask, “Does self-service BI work at all?” The quick and hopeful answer is “Of course, it must work! All my software vendors are pitching self-service platforms to me—this is the future!” Yet sometimes software inventions get ahead of their markets’ needs. Early adopters aren’t always rewarded. Is this the case with self-service BI?
What is software?
Consultants are well known for initially answering every question with a single pat answer: “It depends.” Before you click away from this article, I do have a better answer than that! But in a sense, the success of your self-service BI program does depend—on your expectations. What do you expect self-service BI to provide to your organization?
The software industry delivers amazing innovations. Great software is sometimes indiscernible from magic. So much so that we’ve come to expect magical results from our software programs. But often software is merely a tool to help us do more efficiently what we once did without it. In data analytics, are there really things we can do with software that were impossible with paper and pencil (even if it might have taken 100 times more people and far more time)?
When BI software is viewed as a tool to provide task efficiency within one’s own sphere of experience, it begs the question whether a self-service BI tool can truly enable every end user to be his or her own data warehouse designer, data scientist, data visualization guru and statistician. Or is that too much to expect?
In the case of self-service BI, the tools are much more approachable and require far less technical expertise to use than the tools we used 10 years ago. It’s game-changing that today we have tools that the average person—with only basic computer skills—can use to gain insights from their data. This is an amazing, often inspiring part of self-service BI. Most every decision can now be based on statistical rather than anecdotal data, and that’s powerful.
But we have to keep our expectations in check. Intuitive software makes self-service users more productive. It enables them to “do analytics” in an hour rather than not having enough time to analyze data at all before making many decisions. It helps them meet their routine needs without queuing up for IT or data scientist assistance. This pervasiveness and accessibility of data is the highly visible impact of the self-service BI revolution.
Complex problems are still complex
This is where we get back to expectations. As Javier points out in his article, data modeling can be complex. Many data modeling problems do indeed exceed the skill capabilities of many self-service users. Self-Service BI tools make manipulating software simpler, but so far have done little to make complex data relationships and statistical problems easier to understand and model in a digital format. Data cleansing, statistical methods and process automation are other areas where self-service users may at times need the help of others with advanced skills.
Data science and computer science experts who spend 4-6 university years followed by a lifetime learning how to solve complex data problems will remain in high demand for the foreseeable future. In fact, such individuals are only seeing more demand for their skills as data analytics techniques become more widely used.
Having helped design dozens of BI initiatives that included self-service components, I’ve seen first-hand that self-service BI really is worth investing in—for most organizations. Line employees and managers won’t remain competitive with their peers forever if they don’t increasingly use data as part of their decision-making process.
Does that mean every user can meet all of his or her data analysis needs without the assistance of specialists and experts?
Usually not, and this is where current expectations of self-service BI technologies are still not perfectly aligned with business and human resources realities. Software is getting easier to use, but the data problems and business questions we use software to solve aren’t getting easier.
Effective business intelligence initiatives will continue to be a partnership between end-users, subject-matter experts and data experts over the foreseeable future.
Self-service BI technologies are improving this partnership by helping users do more and reduce decision cycle times. Data experts are enabled by self-service too—now able to focus less time on routine tasks and creating final reports, and more time improving the quality and comprehensiveness of the data their end-users use. But the partnership is evolving through self-service—not being dissolved by it.
At BlueGranite, we assist business and technology leaders interested in crafting Self-Service Analytics and Reporting strategies. Our flexible approaches can support the initiative by targeting a wide range of supporting activities, from building data repositories specifically crafted for Power Pivot use, to end user training, mentoring, and adoption programs. Be sure to contact usif you are in the process of deciding which flavor of Self-Service Analytics is most appropriate for your organization.