We live in a business intelligence (BI) and analytics world, where people get excited by shiny new objects which we hope and believe will fix all our problems. This blog today considers a couple of those shiny objects, why they may or may not solve all our woes, and also how we humans can still play a role in one of the latest trends: augmented analytics.
Shiny New Objects, Example 1: Data Discovery
Data discovery tools certainly caught the eye of business people a few years ago, who believed that self-service was their new information utopia. They were able to employ user-friendly tools that enabled the uploading and manipulation of personal data, and then the visual exploration of these custom data models. To some degree, they have been successful, whereby business people have been able to extract insight and value from data that was not previously possible.
Over time several challenges have emerged, and it has taken two to three years of actual usage to identify them:
- Errors, leading to bad decisions. Self-service data preparation and content creation can often lead to human error in the data manipulation and analytics creation process, and decisions can very easily be made from erroneous insights. Even worse, analytics and insights are often shared and the errors compound upon themselves
- Silos. Analytics silos are typically formed by self-service users, who prepare their own data, create reports and charts, and share them within their own work group. They tend to exist in their own bubble, creating their own library of analytical content, with their own new calculations and preferences. These silos emerge in other parts of the organization, and it ends up being like the Tower of Babel, where there are multiple analytical groups and multiple versions of the truth
- Scaling. Self-service data discovery usually starts life in a department or smallish work group, where a business leader has elected to invest in a subscription for his/her team. If, in their view, they have been successful and recommend it be rolled out across the entire organization, a very large scalability wall appears. Tools are simply not platforms, which inherently possess all the underlying data/content/adminstration/user/security management that is critical to support larger user audiences.
For businesses and users to discover effective analytics and actionable insights, they need data they can trust, and an environment that can grow with them as needed. Data discovery and self-service needs to be an extension of a well organized, governed, and trusted data strategy and platform. With more data sources available for exploration and opportunity, data quality and data governance become critical best practices. I would say that Information Builders is unique in that we offer a highly scalable BI and analytics platform (WebFOCUS) that has built-in data preparation and metadata management that drives governance, as well as sister families of products dedicated to data integration and data quality.
Shiny New Objects, Example 2: Augmented Analytics
There are several terms that have hit the BI and analytics mainstream of late, including data science, machine learning, artificial Intelligence (AI), and augmented analytics. To me, the biggest impact of these capabilities is the tectonic shift from an all-human analytics world to a semi- or fully automated world.
Regardless of where you stand ethically about human effort being displaced by AI and algorithms, the fact that there are profits to be made means organizations will experiment and adopt these capabilities. Augmented analytics is a term that has caught on as it relates to how humans can be assisted in their quest for data insight and the opportunity for smarter business. Most industry analysts consider the use of machine learning and natural-language generation as primary mechanisms to leverage computers doing what they do best - namely crunching huge volumes of data, identifying patterns, learning from them, and generating recommendations to assist the human user.
Forgive a slightly cynical view of this, but it will not be long until humans are removed from many business equations where analysis and decisions are needed. The concept of augmented quietly shifts to automated, and many data analysts might want to update their resumes. Maybe that is a little dystopian, and I conveniently forgot to mention that human creativity, instinct, and intuition are difficult to displace with rules and code.
There is certainly a place for AI and machine learning, but be aware of the shiny new object syndrome. Augmented analytics is a credible space, with an increasing number of use cases delivering real business value, but humans still have a major role to play.
What About Human-Based Augmented Analytics?
There is a new collaboration function within WebFOCUS, called Intelligent Objects, which enables users to leverage each other's analytics. The secret sauce is a new dynamic metadata layer that is automatically generated as analytics are created across the organization. Based on security privileges, any user can see and access similar and related preexisting views, enabling employees to build on each other’s work, increasing the pace of innovation across the organization, and reducing redundancy.
For example a person analyzing a customer process can drill into a distribution visualization that another employee created, in another department, using a different data set. And the content creators never even have to talk to each other to make it happen.
WebFOCUS offers plenty of AI-related support for data scientists working with Python and R, but this really is another form of augmented analytics that you should employ. It comes out of the box, and leverages perhaps the biggest BI and analytics investment you currently have, your existing (and future) analytics content. This type of augmented analytics has never been done before, and it's available now, with WebFOCUS.
I love the speed of innovation in our industry, but don't be blinded by the shine!
Thanks for your time today.