Deriving Competitive Advantage From Data And Analytics

Greg Firestone, Vice President of Data Science., Allstate

It’s no secret that companies across all business sectors are investing heavily in data and analytics. The rapidly accelerating trend will pay future dividends in the form of increased efficiency, new products, innovative customer experiences, and more. According to a recent study, 53 percent of companies were using big data analytics in 2017, up from 17 percent just two years before.

Allstate recognized the value of data and analytics long before it was a widely established norm. In 1939, we became the first insurer to tailor auto rates by age, mileage and car usage— all key data elements still used in pricing today. Since then, Allstate has strategically accelerated transformation across the enterprise by researching and applying advancements in big data, artificial intelligence (AI) and proactive personalization to business opportunities.

Over the last few years, we’ve invested heavily in our data and analytics capabilities, including bringing in new talent, developing technology to enable our analytics solutions and building programs to create a data-driven culture across the company.

"Leadership support and buy-in is what makes execution happen, and will allow those companies who are doing it well to pull ahead of their competitors"

Deriving competitive advantage from data and analytics is not just about raw talent horsepower—it’s about successful execution achieved through effective partnerships between analytics and business teams, coupled with strong leadership endorsement. Leadership support and buy-in is what makes execution happen, and will allow those companies who are doing it well to pull ahead of their competitors.

A robust, multi-faceted approach to developing the data and analytic function at Allstate supported the formation of these partnerships. For example, we hired business consultants, project managers, and communications experts to support data science teams and demonstrate the potential of their findings. This, coupled with clear alignment to business goals helped us fully leverage data and analytic capabilities to influence business decisions at the ground level. Most importantly, we’ve made large investments in our people through training and development programs.

We’re already beginning to see our work have an impact on the business, all because of a close partnership between our team and the business partners we support. Examples of tangible business impact include retention-improving customer experiences built upon data-driven insights, and the use of machine learning (ML), AI, and natural language processing (NLP) to deliver faster, easier claims processing. Notably, our QuickFoto Claim initiative allows customers to submit collision photographs to adjusters digitally, eliminating the time previously used to travel to in-person inspections. It’s just the beginning, with many advanced analytics projects in flight to help adjusters resolve claims even faster in the near term.

Through our work, we’ve learned some valuable lessons. The following are a few tactics we’ve found to be successful in earning business leaders’ confidence in data and analytic advancements and adoption of our work.

Listen Carefully to Understand Business Partners’ Problems

Before proposing cutting-edge, transformational ideas to business partners, it’s important to listen to their needs first. Falling in love with your own ideas bears the risk of overlooking business partners’ ideas—the very people closest to the root cause of an issue or opportunity. If your data and analytics team is centralized, encourage all exposure to the business unit they support though informational meetings, job shadowing with front-line employees, customer service representatives and more. The best perspectives and solutions to business partners’ needs can often be found “walking in their shoes” for a period of time—and it’s time well spent.

Communicate Effectively

Even if you’ve sourced near-perfect data, triple-checked your underlying assumptions and created highly predictive models, data and analytics won’t be a slam-dunk unless you can show their merit to decision makers in your organization. That’s why it is important to sell your work in a way that makes sense to them. Because data science language may not be universally understood, try offering a demo of your work to bring the conversation up a level from the math to focusing on what’s possible. It also ensures that everyone—both the business and analytics teams—are on the same page and able to see what the solution looks like on an even playing field.

Balance Both Short-term and Long-term Results