Have you seen how analytics vendors are promoting and selling self-service? Well, if you look closer or are just starting to research the topic, you will start seeing some mismatches between the sales talk and reality. Obviously, customers will always have some initial expectations that might not live up to the actual product’s capabilities. However, in this little series of articles, we aim to show you that customers aren’t at fault for having such notions. So, let’s begin, shall we?
“It’s a trap!” said Data Literacy
Technology has evolved a lot in the past decade, and it doesn’t seem like that perpetual trend will be put to a stop anytime soon. Naturally, many fields are benefiting from this, and data analytics is no exception. We’ve seen the introduction of many helpful, intuitive, and advanced features from vendors like Qlik, Microsoft, and Tableau. But no matter how well-made the software is, one thing will always remain at its core – data literacy. So, here lies the first issue with self-service. Despite how friendly and convenient self-service sounds, it still requires you to understand the data and its uses. So, no matter how easy-to-use software is, you’ll need some form of data literacy to make sense of everything.
Without the needed experience or training, you can hope to get the necessary data faster to the people at the top – sure, but with lots of dirty data, improper aggregations, or maybe bad visualizations. We’ve observed similar statements from some of our clients, as well.
One of the most common analytic mistakes in self-service is, in fact, selecting chart types based on visual aesthetics and preferences rather than based on the principles of effective visual communication of data. This, of course, can and most likely will make it really hard for your visualizations to tell a story or bring important insights to light. Additionally, those quick summary reports could present a false or skewed picture, far from reality, if the self-service user isn’t data literate. If you’re familiar with Anscombe’s Quartet, then you pretty much get our point. But to summarize, you can have four data sets with nearly identical descriptive data and yet appear at odds when graphed.
Looking at customer reviews and listening to feedback points out another misconception – the notion of self-service cannot just fix your data quality. We can trace this misunderstanding back to the catchiness of the trend’s term once again. However, even with the current state-of-the-art technology, there isn’t a system that can take data from several data sources and magically make it compatible and usable. To make things worse, we should factor in the increasing volumes and variety of data sources. This is why getting the proper figures across an entire organization can easily turn into a moving target, which requires huge investments of time and resources. Some data experts can rightfully argue that it is better to use “roughly right and relevant” data than being in a “too late to make decisions” situation. Furthermore, just using data “straight outta the box” can lead to wrong results and cause issues when making business decisions. This all, of course, ties back to the data literacy problem we highlighted earlier and could be avoided with a little bit of longer-term vision and know-how.
“So, what are the potential solutions for Self-Service?”
With all that said, it’s clear that if companies wish to adopt self-service, they should invest first and foremost in data literacy. Ensuring all your employees are adequately trained would likely avoid most of the missteps we mentioned earlier. However, don’t just stop there. Choose your data champions and allow them to attend crucial meetings with top-level decision-makers to avoid resistance to this transition to a data-driven culture. Also, to confront potential data governance and data quality issues, you need to allow leadership to be in charge of the vetting process. Taking measures and preparing for the unexpected could spare you from the pitfalls of an overhyped trend, and as you can see, most of the issues can be solved with a little know-how.
Conclusion
Having lots of business users creating their own reports, not using help from the IT departments, and providing the necessary data to the right people in a quick manner has undeniable benefits. However, companies need to be careful when adopting self-service. Without reliable data governance and data quality management practices, you can potentially end up with endless iterations of the data at hand. In the meantime, analytics vendors need to consider if it would be more beneficial to specify what self-service is and isn’t, or if it would be better to just relabel the trend to something less misleading to the public.
