The marketers of today know the importance of engaging consumers through personalized messaging with the right touch points. Still, it becomes difficult to cut through the noise of instant email and piles of direct mail. With companies locking in on the necessity of reaching their customers with as much personal detail as possible, how does a brand continue to stand out? That’s where the secret of data science comes in, and marketers are beginning to use this lucrative resource to restructure entire advertising and marketing campaigns.
Data science is a vastly growing field that continues to expand as trillions of bytes of consumer data are collected every day. At its core, data science takes numbers and gives them meaning. This meaning can predict buyer behavior, and in turn influence marketing campaigns and major business decisions to maximize return on investment. Digital Shuffle, CCG’s agency partner, has created a data science platform built on cloud-first native operations utilizing Python and enterprise business intelligence to benefit our CCG customers from the ground up. The beauty of this proprietary platform is that it uses your existing data.
Customer Analytics are our passion. Digital Shuffle uses three distinct behavioral models to help marketers make informed decisions about their marketing campaigns. These include customer churn, associate rules mining, and customer lifetime value. Digital Shuffle collects data from our clients existing customer relationship management (CRM) software and analyzes it to create the three behavioral models. The agency then puts the results into a digestible form for the client, replacing the need for them to have their own analyst.
By definition, customer churn is when a business loses a customer for any reason. Churn can be passive or active. An example of passive churn would be consumer turnover due to a membership expiring that the consumer never reactivated. Active churn would be if a consumer left one company for another because they liked a competitor’s product better. The churn behavioral model assigns a quantitative value to the likelihood of a consumer leaving the company. The results of this type of assessment allow the client to make big picture decisions based on numbers and facts rather than assumptions.
Associate rules mining is a model that predicts consumer behavior as it tracks consumer movements to forecast what a consumer will do next. The model can track purchasing habits of each individual consumer. For example, when a retailer collects purchasing data from consumers, the model can surmise that whenever a certain customer buys a shirt, they will also buy matching pants. With this information, the model establishes association rules which allow the company to tailor a campaign, incentive, or discount to encourage the consumer to continue this purchasing habit.
Finally, customer lifetime value assesses a consumers’ value to a company by ranking them based on the revenue a customer will generate for a business throughout the lifespan of their relationship. With these consumer segments created, a company can modify its marketing campaigns and offers based on a consumer’s ranked value. For example, a business could offer high value customers incentives such as referral codes and cashback offers. Or it could use loyalty programs to entice lower value customers, who are more likely to leave, to re-establish allegiance to the company.
Digital Shuffle recently completed an in-depth data science effort with a customer in the transportation industry. The customer operates in both B2B and B2C markets across the U.S., providing testing and certification that verifies the technical knowledge of service professionals in a variety of disciplines. With more than 50 different certification tests, millions of database records, and over a quarter million active professionals, the customer was keen to have Digital Shuffle dive deep into their customer analytics.
Although the customer actively communicates with their consumers via omni-channel communications, the concept of using artificial intelligence and proprietary algorithms to identify purchasing patterns and probabilities within their consumer data was very exciting. Digital Shuffle’s data science platform allows customers to identify and slice their consumer population based on location, age, education level, employer types, years of certification, number of certifications, purchase behavior, and more, to allow for ultra-personalized targeting in marketing efforts.
Digital Shuffle began by creating a customer churn model, specifically designed to mitigate customer churn occurring from non-recertification (a professional certification is valid for a period of 5 years, then a recertification test must be taken to keep certifications up to date). Digital Shuffle analyzed a data set comprised of 43,108 unique individuals that had pending certifications expiring in 2021. The model was able to predict that 12,435 of those individuals were in a “high churn” category, 25,168 were in the “moderate” risk of churn category, and 16,697 were in the “low-risk” category. Accordingly, Digital Shuffle created a series of multi-touch, multi-channel communications with varying incentives (relative to churn probability). The following explains the paths taken to market to each churn category:
Low Risk of Churn: Multi-touch email communications deployed over a 45-day period, focused on communicating when it was time for renewal, why renewal is important, and where to renew.
Moderate Risk of Churn: Multi-touch email and direct mail campaigns deployed over a 45-day period, focused on communicating that it was time for renewal, why renewal was important, a variable incentive to renew, and where to renew.
High Risk of Churn: Multi-touch email and direct mail campaigns deployed over a 30-day period, focused on communicating it was time for renewal, where to renew, why renewal is important, an increased variable incentive to renew, and why other similar professionals would choose to renew.
A parallel effort was executed to incorporate an associate rules mining model, also known as “cross-sell/up-sell”. Data science provides answers to the question, “When a customer previously purchased X, what are they most likely to do next, and when are they likely to do it?” Armed with this information, Digital Shuffle executed a series of omni-channel print/web/email campaigns to suggest subsequent certifications to take the next step on the professional’s career path. Leveraging the power of data science, Digital Shuffle created completely unique suggestions for over 200,000 individuals, all completely tailored to the unique purchasing habits of each individual.
Data science has completely changed the way Digital Shuffle’s customers identify, understand, and engage their past, current, and prospective consumers. The ability to leverage extremely complex algorithms and artificial intelligence to visualize and predict consumer behavior is revolutionary. The days of basic demographic, geographic, and psychographic data are truly in the rear-view mirror. Interested in learning more? Visit DigitalShuffle.com, or contact your CCG rep to schedule a discovery call.
By Marley Niesz