Earlier this month, Gartner published their annual Top 10 Data and Analytics Trends. Of course, non-technology trends are excluded by definition from the tech list. Yet the use of soft skills—data literacy, storytelling, and data ethics—came up throughout the blog post. While predictive modeling captures the imagination and the headlines, more often than not, soft skills have made the difference in successful predictive analytics projects.
Three soft-skill themes are especially important in predictive analytics. The first: predictive modeling requires business and technical perspectives throughout the process. Second, the widespread use of predictive analytics requires more people who can communicate fluently about data and analytics. The third deals with the use of storytelling to convey the value of predictive analytics throughout the organization.
The best predictive analytics projects involve non-technical roles
Trend No. 1 and Trend No. 2 on the list both deal with augmented analytics: augmented analytics and augmented data management, respectively. This Syntelli blog post discussed augmented analytics. Essentially, augmented analytics uses techniques to automate some of the heavy lifting of analysis, so people get insights faster.
One of the important benefits of this trend is that it invites more business roles into the predictive analytics process. Augmentation even helps with the more technical aspects of predictive modeling, like splitting data sets for testing, validation, cross-validation, and other data preparation steps that support predictive modeling directly.
The reason why the involvement of non-technical roles is so critical is that predictive modeling is not a technical process. Predictive analytics really automates the business process of decision making. It just happens to use data. Predictive analytics separates from its descriptive and diagnostic cousins by emphasizing prediction’s role within business operations. Without business involvement early and often, projects bog down in technical details without the benefit of alignment with the business needs.
As the use of predictive analytics grows, data fluency will too
Trend No. 1 in the Gartner list specifically mentioned the importance of data literacy—”the ability to read, write and communicate data in context”—because it is one of the most important skills in organizations.
Data literacy for predictive analytics has had precursors: literacy and a data culture have grown in organizations that adopt lean and six sigma methodologies. People in the cultures speak the language of measurement, which is critical for fast exchange of ideas across functions.
The same is true for predictive analytics and predictive modeling. The literacy metaphor is apt because people who speak the language of data have an advantage, and the organization with a greater data literate rate is simply going to be more productive. Soon, the language of business will be a language of data.
Stories clarify insights derived from predictive modeling and motivate people to act
Ultimately, the power of predictive analytics is the power to act more intelligently. Gut decision-making gives way to a data-driven enterprise. For people to support the business processes around predictive modeling, you need to ground the technical aspects of modeling in the real world.
What does a prediction really mean for a marketer writing content? Should the marketer write copy more in tune with a certain persona, when the model gives unique confidence scores to individuals?
Gartner provides another example with predictive modeling in Trend No. 7, Explainable AI. Gartner states, “Without acceptable explanation, auto-generated insights or ‘black-box’ approaches to AI can cause concerns about regulation, reputation, accountability and model bias.” In other words, certain business imperatives, like fairness in lending practices, simply require transparency.
Storytelling is the ability to communicate complex predictive analytics insights using a ‘narrative arc’ that is more accessible to people. Gartner refers to an engrossing article about Florence Nightingale, mortality rate, and the Crimean War to demonstrate how—in the mid-1850s—Nightingale’s creative use of visualization made understanding data easier. Storytelling helps people engage with the results of predictive analytics as well.
Syntelli Solutions provides expertise in predictive analytics and predictive modeling. Contact us to learn more.