Low-code tools are revolutionizing business, enabling citizen developers to create new business applications that drive innovation. Now, the same thing is starting to happen for citizen data scientists.
AI has taken center stage as more companies across industries shift to become “data-driven.” But becoming a data-driven organization requires knowing where your data comes from, the best ways to analyze it, and how to communicate its value effectively. And the reality is that most organizations still struggle to attract and retain talent with the skills to do that effectively.
Low-code tools have the power to help. Expanding the use of these technologies and training more employees to use them can make a transformational impact and help organizations scale data science across business units. Yet too few organizations are off and running in the low code journeys.
Here are six key reasons that business leaders should consider taking that first step to build data science capacity throughout their organizations.
Subject Matter Expertise – Too often, the people within an organization who have access to data analytics tools aren’t particularly close to the problems that the data may help to solve. We can’t expect data scientists to be experts in finance, marketing, customer service, manufacturing, logistics, and all of the other departments and functions that make a company run. Even if they have access to data from different lines of business, data scientists may lack the subject matter expertise to always put it to good use. One way to solve this is to simply reverse the equation – bringing one new skill (data science) to departments across the enterprise, rather than expecting data scientists to master every other aspect of the business.
Diversity of Ideas – When more people in an organization have data analytics skills, they can try out a greater number of ideas – increasing the likelihood that they’ll land on valuable insights. Often, the answers lie in surprising places. One car manufacturer that I’ve worked with used data science to help business leaders figure out the source of manufacturing delays. A number of stakeholders in the company had assumed that the problem lay with the complexity of a certain vehicle’s power train. But in fact, analytics revealed that the delays were actually being caused by a seemingly minor change in an accessory package.
Faster Time-to-Insight – For most organizations, the time between someone having an idea about how to solve a problem with analytics, and actually being able to test that idea out, is far too long. The COVID-19 pandemic showed just how quickly conditions on the ground can shift, and businesses need to be able to quickly arrive at data-driven insights to adjust before they’re left behind. Data science teams are typically already overloaded with projects, making it difficult for them to respond rapidly to new developments. But by creating analytics capabilities throughout the organization, companies can tighten up their analytics timelines.
Awareness of Potential Problems – Mark Twain once said that there are three kinds of lies: “lies, damned lies, and statistics.” The idea here isn’t that data is never helpful, but rather that statistics are sometimes used to support faulty conclusions. (If you confuse correlation for causation, you might decide that carrying an umbrella causes it to rain.) Most people without analytics training aren’t familiar with common data science traps. Have you ever heard of Simpson’s paradox, for example? Have your employees? If not, it may be difficult for stakeholders across your organization to spot problems with how analytics are being used.
Improved Collaboration – The past few years have shown us just how important collaboration is to productivity. Collaborative authoring tools like Google Docs, for instance, have completely upended traditional workflows. (No more emailing different versions of a document around and losing track of which one is the latest.) And especially during the pandemic, video collaboration became central to many companies’ operating models. Collaborative no-code analytics tools aren’t yet widely available, but these solutions are currently being developed. Once they’re ready, they will help stakeholders across different business units work together in real-time to collaboratively solve problems.
Building a Talent Pipeline – When it comes to data analytics talent, you can either compete for it, or you can create it. By training up people who are already on staff, companies ensure that they have ready access to data science expertise. Also, longtime employees who acquire analytics skills may be more likely to stick around than the data science experts who will continue to be bombarded by offers, even after they are hired.
The idea of becoming trained in analytics may at first seem intimidating to line-of-business employees. But remember: The idea of getting trained in Microsoft Excel was once intimidating to people. Today, basic computing skills are simply part of the job for nearly all knowledge workers. In the coming years, we may see the same thing happen for data science. But first, companies must invest in education. With the proper training and development, we companies can create a robust stable of citizen data scientists who can quickly and easily do the analysis they need to drive business forward.
Tim Kraska is an Associate Professor of Electrical Engineering and Computer Science in MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), where he co-leads the Data Systems Group and is part of the Systems Community of Research (CoR). He is also the Founding Co-Director of MIT Data System and AI Lab (DSAIL). His research is focused on dramatically increasing the efficiency of data-intensive systems and democratizing data science through machine learning (ML). Kraska is also lead instructor of the new MIT Professional Education course, “No-Code Analytics and AI”, which debuts in 2022.