As decision science carves a larger niche for itself within the field of data science—which has seen vast growth as a discipline—we’re seeing a shift in the way teams use the wealth of data available to them. Data analysts and reporters have been acquiring, organizing, and cleaning data for years. Now, with better reporting and collaboration, data scientists can take that data, analyze it, and use it to make more informed decisions.
While that current structure works, it’s continuing to evolve and improve. Machine learning now enables companies to move from a reactive, cause-and-effect mindset to a more predictive one. Machine learning and deep learning models get better at predictions every day, but the decision to focus on the future must still come from within a data science team. With that need in mind, companies are investing in teams with the skills to take available data and apply it to potential future situations.
At Clearlink, our evolution from a reactive decision science team to a proactive one has been a natural progression as our team has evolved. As we’ve continuously pushed progress in data capabilities and added key talent, we’ve been able to stay dynamic instead of getting stuck in one reactive mindset. The positive results we’ve seen reinforce that making the shift from reactive to proactive is worthwhile, wherever your data science team is at currently.
Start with a Strong Foundation
Clearlink’s first foray into data science consisted of a small team. In just a few short years, we’ve grown significantly and now enjoy a close relationship with the marketing and sales divisions of the business. When we started, however, we needed to establish a framework that would allow us to react to real-time needs before we could even think about being proactive. That meant building a vast data warehouse over time. As we built the warehouse and gathered data, we were able to start reporting, learning, and reacting.
This type of reporting was certainly valuable—being able to understand and react to past customer behavior helped us improve the experience for future customers. Unfortunately, this still left current customers with experiences that we couldn’t change in real-time. By looking back instead of forward, we gained good insights and direction, but we left customers feeling frustrated in the moment. Using more advanced analytics, we’ve been able to start predicting customer journeys before they happen, which enables us to help people in real-time rather than after we see them abandon.
Proactively Change Customer Experiences: Traffic Control
A proactive approach has specifically helped us with balancing call traffic and marketing campaigns. Typically, when one of our sales teams gets overwhelmed with call volume, that level of volume triggers specific PPC campaigns to shut down. This limits the number of customers being driven to the call center—only to experience long wait times and potentially dropping out of the queue—and prevents our marketing team from overspending on campaigns we don’t need. This process identifies when traffic is at capacity and helps us scale back accordingly.
Historically, this process involved looking at a specific point in time and gauging our situation. If we looked at a specific time and saw a drop in calls, we would increase the number of PPC ads we were running to increase call volume. Five minutes later, however, calls could spike again. Our sales agents would be overwhelmed, and we’d be shutting down those same campaigns. This approach works, but it too often involves looking backward and reacting instead of thinking ahead.
Now, machine learning can anticipate those fluctuations in call volume and help us avoid overreacting. Based on data we already have, we built models that show the call volume we should expect in the next 15 minutes, as well as how many agents should become available in the next 15 minutes. That allows us to plan ahead about how to balance agent availability with marketing campaigns, giving us time to make strategic choices backed up by data in advance.
With a more proactive approach, we can offer a more intelligent customer experience by managing call volume more consistently. This smooths the customer path and, as a result, strengthens our brand. When customers benefit, the business benefits. Further, this approach facilitates more strategic spending. We shut down campaigns that don’t result in added revenue, and we decrease our customer abandon rate—allowing us to optimize our investments in the areas that are going to help us most.
Encourage a Proactive Approach
If your team is still in a reactive mindset, encourage them to think differently to pave the way for a more proactive approach.
- Use case studies: Show your team how machine learning and predictive analytics have helped other companies. If possible, create hypothetical case studies illustrating the way proactive thinking could help solve a problem for your team.
- Talk about what the customer goes through: Closely outlining exactly what your customers experience when interacting with your brand is a great way to illuminate new issues or help your team see areas to improve that they might have previously missed.
- Show your team potential (or real) impact: Data science teams have a love for analytics, but their motivation ultimately comes from the impact those analytics have. Motivate them by identifying opportunities for your team to make a difference and then showing them positive results for the bottom line and for the customer.
At Clearlink, the traffic control example is part of an AI layer that underlies our customers’ journeys, which is aimed at making every step informed—pro-actively—for the customer, based on everything that happened beforehand. This real-time ability to impact people positively through data science has really inspired our decision science team to embrace a more proactive mindset. It has a significant impact on our bottom line and it impacts both the business and the customers positively, which continues to be further motivation to stick with this new way of thinking.
If your team is struggling, keep in mind that the transition doesn’t need to happen overnight. Build the right foundation, strive to use data to make a difference in the business, and aim to start thinking and planning ahead. Your team will follow your lead and start moving in a proactive direction, too.
Latest posts by Pieter van Ispelen
- Moving Your Decision Science Team from Reactive to Proactive - July 9, 2018