A propensity model calculates the likelihood of a prospective or current customer’s next steps. Understanding likely next steps helps companies deliver better experiences, increase loyalty, reduce churn and build value for the organization.

Read more

Customer Experience Tools and Trends

Our insights share how the right CX tools make customers’ lives richer and more fulfilling and strengthen a company’s economics by holding down costs and securing new revenue streams.

How companies use propensity models

• Churn prediction. Precise modeling raises warning signs of churn and allows a company to take measures to prevent churn before it happens.
• Product recommendations. If companies—particularly online retailers—know which products and services their customers will likely buy next, they can provide relevant recommendations to help speed decision making.
• Offer and product customization. Companies can boost sales and improve loyalty by offering the products and services that are most relevant to a specific customer.
• Calculating customer lifetime value. The ability to forecast a customer’s total spending over time can help companies prioritize segments and evaluate potential investments in product features and new value propositions.
• Predicting Net Promoter Score℠ outcomes. Data analytics can help companies identify which customers will likely become promoters, passives or detractors without surveying them. A predictive Net Promoter Score enhances service and recovery, reduces survey volumes and associated costs, and helps companies prioritize their customer experience initiatives.

Key considerations  

• Put business science before data science. A company’s predictive analytics goals should reflect its broader ambitions. Leaders identify the decisions they want to make and the data they will need.
• Design the analytics with the last mile of adoption in mind. Bring data scientists and business stakeholders together to create the best solution. Make it easy for the front line and decision makers to adopt the output.
• Look beyond traditional analytics. Go beyond structured enterprise data and consider sources such as social media, web scraping, customer interaction transcripts, log files, sensor data, images and other publicly available content.
• Find the shortest path from insight to action. Companies should maximize the yield on their existing data and become nimble enough to act on insights quickly. After all, the value comes from the action and not the input.
• Test, learn and iterate. Take an Agile approach to continuously improve the solution and outcomes. Pilot new solutions on real customers and processes, even if the tools are barely viable, rather than wait for the seemingly perfect solution.
• Manage the analytics transition with purpose. Successful predictive analytics solutions require sound operating models to bridge strategy and execution across departments and functions.