Predictive analytics – defined as “the use of statistical algorithms and machine learning techniques to identify the likelihood of future outcomes based on historical data (Source: www.techemergence.com)” – is being used by marketers across industries. After, Inc. uses predictive analytics in its Warranty Marketing Solutions to drive high response rates, attach rates and ROI. The top three modelling techniques we use most often include: customer profiling, propensity models, and customer lifetime value.
1. Customer Profiling / Segmenting
An ideal customer profile (ICP) is a snapshot of your ideal customer(s), also known as a “buyer persona”. ICPs are built using shared attributes of your top customers (or those you’d like to target) and can include gender, income, age, marital status, number of children, job title, and hobbies/interests. ICPs also include purchase motivations, technology use, and online behaviours in order to help guide messaging and advertising strategies.
Figure 3: Customer Profile: “Data Dave”
For prospects with a “Data Dave” profile, the most relevant messaging for a refrigerator manufacturer might focus on how an extended warranty can save thousands of dollars in costly repairs. The manufacturer might also include testimonial videos that speak to customers’ real financial savings.
How this model optimizes marketing spend: Developing detailed ICPs and segmenting targets with similar characteristics helps guide warranty marketers to use the most relevant messaging, imagery, and marketing tactics for each group. Customer segmentation drives more personalized marketing – and the result is increased response rates, sales, and ROI.
2. Propensity Models
A propensity model, or “likelihood to purchase” model, predicts which prospects are most likely to purchase an extended warranty product. It scores prospects based on how closely they resemble customers who have purchased warranty products in the past.
Propensity models are built from large data sets that contain demographic variables, such as age, gender, education, etc., as well as transactional variables, such as product, price, marketing channel, and history of previous extended warranty purchases. A statistician will run various analyses to determine which variables are highly correlated with purchase. The result will be a model that compares the variables or attributes of each prospect against the customer data set and assigns them a “likelihood to purchase” score.
How this model optimizes marketing spend: Propensity models are extremely useful in warranty marketing because they group prospects into tiers based on high, medium, or low likelihood to purchase a warranty. Marketers can target the highest propensity prospects, instead of marketing to the entire prospect list, thus minimizing spend and maximizing ROI. In addition, marketers may decide to send the “high propensity group” multiple offers across channels, because statistically they are more likely to purchase and therefore worth the higher marketing spend.
3. Customer Lifetime Value (CLV)
The third type of model used in Extended Warranty Marketing is Customer Lifetime Value (CLV). CLV is “the total amount of money a customer is expected to spend in your business, or on your products, during their lifetime (Source: Shopify).” This is an important number for warranty marketers because it represents an upper limit on what businesses should spend to acquire a new customer.
For example, let’s say that Data Dave’s persona – although expensive to acquire – spends a lot with the company over his lifetime. Not only does he purchase the initial extended warranty, but he purchases additional accessories and becomes a repeat buyer of this manufacturer’s refrigerator and warranties. Therefore, “Data Dave” will have a high CLV.
How this model optimizes marketing spend: Being able to accurately predict the future dollar value of a set of customers is an incredible asset. However, CLV should always be used in conjunction with propensity scores in extended warranty customer acquisition strategies.
Using these three predictive modelling techniques in warranty marketing campaigns provides better forecasting, optimized budgets, minimum acquisition costs, and maximized ROI.
To read a case study of how After, Inc. used predictive analytics and multi-channel marketing to more than double sales and profits for Electrolux North America’s Extended Service Agreement (ESA) Program in the first year, click here: https://afterinc.com/electrolux-case-study/.