A new collaborative book project from the makers of leading semantic layer platform, AtScale, offers collective advice from a host of specialists from data science and business intelligence. Contributors include Bill Inmon ‘The father of the data warehouse’, Kirk Borne ‘renowned chief data scientist & AI influencer’, Juan Gorricho, VP of Global Data & BI at Visa, Ram Kumar, CDAO at Cigna, Megan Brown, Director of Data Literacy at Starbucks, and a dozen more.
The book – Make AI and BI Work At Scale – comprehensively covers topics from aligning analytics to business outcomes to ethics in AI. Along the way, readers will pick up insights on concepts central to the journey to full data maturity. This includes data warehousing – and its evolution into “data lakehousing”, the “semantic layer” which adds meaning to data, and challenges around building a culture of data literacy within an organization.
If the global pandemic has taught us anything as business leaders and strategists, it’s that understanding how these topics fit with an organization’s digital transformation is key to building resilience and readiness. All too frequently, the difference between a business that has thrived and a business that struggled or sank was its ability to leverage technology and, specifically, data, to adapt to changing circumstances.
In this piece, I will take a sneak-peak at some of the important lessons from industry leaders that I picked up from the book. Some of them are very high-level, strategic tips applicable in some form to just about any organization, whereas others may be more specific to particular industries or challenges. All are valuable and make a great jumping-off point for gaining a deeper understanding of the philosophies and methodologies of these thought leaders.
Sometimes instinct is the only option. At all other times, use data.
Prashanth Southekal, Managing Principal of data analytics consulting and education company DBP-Institute, explains that understanding when to rely on existing knowledge or even gut instinct and when there’s a need to collect and analyze data is an essential part of the puzzle. Firefighters and intensive care doctors, for example, may find themselves more reliant on their experience and instinct when acting in emergency and crisis situations. Similarly, instinct is valuable in situations where historical data is not available, such as when an organization makes a move towards a new market that hasn’t been tapped before. However, when dealing with complex hypotheses with many interrelated variables, particularly when their relationship is not well understood, instinct is unlikely to get us too far. “Most adults can store and process between five and nine variables in their short-term memory,” Southekal reminds us, “So with complex situations with many interdependent variables, one needs to rely on data and the computing power of machines … to derive insights. Holistic insights are a combination of both intuition and data analytics.”
Business intelligence (BI) and data science (DS) – The yin and yang of value creation
There’s often confusion between the dual disciplines of BI and DS. Both deal with information and insights, leading many people to think they’re essentially the same! Ram Kumar, Chief Data and Analytics Officer (CDAO) at Cigna -International Markets reminds us that this isn’t the case: “I call BI and DS the “yin” and “yang” of data-driven value creation … because they complement each other and each can solve different business problems.”
Five catalysts for improving business outcomes with data, insights, and analytics
According to Brian Prascak, Chief Insights Officer at Naratav, big data, cloud, advanced analytics, modern data platforms, and data virtualization are the key elements of data strategy within today’s most forward-looking (and, therefore, data-driven) organizations. Big data is the fuel, cloud technology centralizes the toolsets we need, analytics supplies the learning, and virtualization allows us to apply the “semantic layer” that gives meaning to data. Understanding how these elements work together is the secret to unlocking the true value of digital transformation in any organization.
The semantic layer drives data democratization…
So how does the semantic layer add meaning? Well, as with the concept of the “semantic web,” it revolves around understanding the relationship between different elements of a dataset, rather than just the values of the data itself. In this way, the same dataset can drive the creation of different types of value across different teams or groups of data users. As Kirk Borne, Chief Science Officer at DataPrime puts it, “Perhaps the greatest achievement of the semantic layer is to provide different data professionals with easy access to the data needed for their specific roles and tasks.” The addition of a semantic layer to a dataset allows data scientists, BI analysts, and other enterprise data users to spend less time gathering and preparing data and more time generating the insights that matter to the business.
…As well as digital transformation!
Slickdeals is the largest external referrer of traffic to Amazon, eBay, and Walmart. Since starting out in 2012, the business has undergone a top-to-bottom journey of digital transformation, evolving from an entirely on-premises stack to a fully modernized, hybrid data architecture. This means that users across the entire business can now harness the power of data and analytics to make decisions that are crucial to their job roles. “Our goal has always been for employees across the company to easily access data and gain business insights without needing to learn to code or forcing them to use a specific tool … we’ve always made it our mission to make data accessible to people when and how they need it,” writes Greg Mabrito, Director of data and analytics.
BI is about looking backward, and DS is about looking forward
A key and simple differentiator that can help us immediately understand the roles that these two disciplines play. BI lets us know what is happening in our business and how we can interpret historical data to learn how different factors have affected our performance. DS, on the other hand, is concerned with the future. “Businesses need insights on future performance, and that’s where data science comes in,” writes Narendra Narukulla, Denior Data Science Engineer at Wendy’s. And they don’t always rely on exactly the same datasets – around 40% of enterprise data science projects are completed using data that’s been prepared by BI teams.
Combining BI, DS, and the semantic layer is the way forward
Understanding the past and predicting the future is key to driving growth and success in today’s world, where every company is quickly discovering the value of becoming a tech company. Whereas BI and DS may previously have been considered siloed disciplines, the value of bringing them together and adopting a unified approach to data and analytics, thereby enabling the creation of a semantic layer, is increasingly apparent. According to AtScale founder and CTO Dave Mariani, organizations that are truly data-driven combine descriptive (what happened?), diagnostic (why?), predictive (what will..?), and prescriptive (what should..?) analytics with the abstraction of the semantic layer. This unlocks the true power of data and prepares an organization to prepare for the future from a position of knowledge and strength.