Business intelligence has shifted beyond the experimental stages to become a critical aspect of business evolution. Previously, the burden fell on IT to provide the data that drives decisions. The growing pains were many, as legacy architectures typically were not designed for data sharing among disparate systems and departments.
Today, modern systems are designed with BI in mind, and most IT departments have revised and upgraded their systems to handle this new dimension in information delivery. With these newfound troves of well-organized data at their disposal, businesses are adopting new solutions to provide individual stakeholders with the ability to produce their own insights.
BI analysts and data scientists have historically straddled a middle-ground. They attempt to empower the end user while also producing reports for business decisions. But as BI evolves into a self-service discipline, does this mean the BI department will go the way of the Dodo?
Developing Standards for Data
Although it seems like everyone is ready to get their hands dirty with data, there are inherent realities to this democratization. There are many ways to accidentally (or intentionally) misinterpret data. Flawed correlations, data dredging, misleading visualizations, or small sample size can distort perceptions to the untrained eye. Tyler Vigen’s data fallacy site Spurious Correlations demonstrates how seemingly sure-fire data discoveries can actually be completely random (and pretty funny).
Without proper standards and governance, each department within an organization is free to make their own discoveries. However, if these insights do not align across departments or are based on faulty assumptions, organizational dystopia can ruin the best of intentions. Pinterest is famously guilty of using data to tread on user privacy, making the false assumption that many of its users were engaged to be married. (They weren’t.) The resulting marketing campaign not only damaged the site’s credibility but also exposed some underlying privacy vulnerabilities.
The Latest Definition of BI
Although the term “business intelligence” was first coined in 1865, modern business intelligence is still an evolving field. The current rage in BI is new tools such as Tableau and Power BI. They enable professionals with varying skill levels to do their own analysis.
Tableau and PowerBI are probably best-known for making beautiful visualizations. They really are awesome. However, those visualizations represent a lot of behind the scenes work. Even the most involved viz is 20 percent design. The other 80 percent is data extraction, modeling, and calculation — which are outside the scope of those in non-technical positions.
Business intelligence will likely remain skewed heavily toward the “back office” work, particularly as the self-service model progresses. After all, once everyone can build their reports, you can get back to the hard skills. There are dangers in this model, however.
Businesses who have adopted self-service analytics are uncovering a new side effect. With individual departments producing their own reports and insights, there is a lack of cohesiveness or big-picture understanding of what all this data really means. This affects the ability of an organization to act with a clear intent on any new data insights.
Can Self-Service BI Replace BI Pros?
With too many hands in the data cookie jar, businesses can be missing key insights, misinterpreting data, and performing incorrect analyses. These examples are central to the issues with self-service BI. Their converse is also exactly what BI professionals can legitimately provide.
The future of business intelligence probably isn’t one where the professionals take their hands off the wheel entirely. It’s more likely that they develop robust purpose-built self-service tools. With guard rails in place, BI professionals can help focus their data analysis — and the outcomes.
The Evolving State of BI
The amount of data that needs to be analyzed is increasing daily. While it’s likely that self-service BI is going to grow stronger, the need for a strong BI team will grow along with this demand. Many employees, when trained on how to create reports from self-service dashboarding tools, are going to find out just how cool data can be.
Even in a self-service world, businesses are realizing BI teams should still be responsible for organizing and presenting data observations. Their analytic skills are needed in order to keep a careful watch for fallacious analyses that could doom a well-intentioned effort. And finally, BI analysts and data scientists are necessary for an ecosystem to remain focused on the big picture.
While BI pros may experience some short-term turbulence, the long term outlook is that they will remain central to data organization, presentation of observations, and big picture oversight. This is going to ensure a higher level of synchronicity in new intelligence-based initiatives.