Customer Cyclical Behavioral Analytics
In a recent Big Data Vision Workshop services engagement, our data science team explored a concept that I find very interesting and immensely actionable – uncovering and exploiting customer time period or cyclical purchase and engagement behaviors. Let’s call this “Customer Cyclical Behavioral Analytics.” Let me explain…
Customers exhibit certain behaviors and engagement patterns that are highly predictable within certain time periods. For example, customers could have behaviors and engagements related to the following time periods or cycles:
- Daily expenditures such as coffee, lunch, newspapers, commuting costs, etc.
- Weekly expenditures such as groceries, household goods, healthcare supplies, gasoline, magazines, etc.
- Monthly expenditures such as credit card payments, rent, lease payments, mortgages, utilities, cable, telephone, etc.
- Quarterly expenditures such as tax payments, insurance payments, etc.
- Semi-annual expenditures such as dentist visits, doctor visits, etc.
- Annual expenditures such as vacations, birthdays, anniversaries, taxes, etc.
- Seasonal expenditures such as Christmas, Thanksgiving, Valentines, Easter, Independence Day, etc.
Identifying and predicting these cyclical repeatable expenditures and engagements can be key to driving a highly personalized and more relevant engagement with your customers (see Figure 1).
Figure 1: Customer Cyclical Behavioral Analytics
My own personal story is related to vacations. Every year for the past 20 years, we vacation the first week in September at the same location…the exact same location including resort. And every year for the past 20 years my credit card company has rejected my first transactions while on vacation…even though I have been going to the same location at the same time of year for the past 20 years. I finally decided to fire that credit card company.
Since many of these cyclical behaviors are tied to product or service categories, one can quickly see the value of creating customer behavioral or analytic profiles at the product and service categories (e.g., grocers, coffee, utilities, financial, credit card, healthcare, household, etc.). Having predictable insights into customer’s product or service category behaviors, propensities and usage patterns can be used in the following applications:
- Personalized marketing
- Customer retention
- Merchant co-marketing effectiveness
- Customer cross-sell and up-sell
- Promotional effectiveness
- New product introduction effectiveness
- Fraud detection
- Identity protection
- Defaults / bankruptcies
Once you have identified a customer’s normal purchase and engagement cycles, then these cycles can be monitored for any “unusual” changes that may present a business opportunity. For example, a family may have a propensity to purchase groceries once every 5.5 days. A change in the groceries purchase cycle from 5.5 days to 2.5 days may indicate a student home from college for the summer (or when my son max comes home from college, a change in our grocery purchase cycle to every 1.0 days!). If the grocery store can detect such a change in the grocery buying cycle, then this provides an opportunity for the grocery store to promote bulk buying to this customer (e.g., save $6 when you buy four boxes of cereal, buy two pounds of sweet potatoes an get two pounds free).
There are also events that are more “one off” events such as weddings (or at least not very often), retirements, high school graduations, college graduations, births, deaths, job changes, career changes and illnesses. These life stage events don’t happen with a cyclical regularity. However the ability to analyze detailed customer purchase and engagement patterns can help to identify or predict when these life stage events may be occurring. Predicting when such a life stage change is occurring can drive a much more compelling, differentiated and profitable customer engagement. But that’s a topic for a different blog.
Organizations can build the detailed analytics at the individual customer and product/service category level to identify, predict and act on these cyclical behavioral patterns. Once you have identified and quantified those behaviors, you can leverage that cyclical behavioral analysis across a number of use cases including marketing effectiveness, net promoter scores, customer churn, fraud, = inventory, production, and logistics.