Traditional Measures vs. Loyalty Builders

Surveys

Many marketers use customer satisfaction surveys to measure the health of their business. Satisfaction surveys have many uses but have no value as predictive tools. Time and again its been proved that there is absolutely no correlation between customer survey results and actual customer behavior.

RFM

Recency of Purchase, Frequency of Purchase, and Amount of Purchase taken together are the most basic tools for measuring loyalty. The diagram below shows why RFM, while better than nothing, is not very effective.

Customer	Jan	July	Sep	Dec	Jan
Customer A	$1500	$1000	$500	$200	$100
Customer B	$100	$200	$500	$1000	$1500

In this example, Customer A and B have bought in  each of the last 4 quarters so their recency and frequency are the same.  The amount of their purchases, $3,300, is also the same.  However, it’s obvious that
customer A’s loyalty is growing while customer B’s is declining. 

Regression Analysis

Regression models are linear (vs. non-linear) and face the problem of making trade-offs among the metrics used composing the model. Loyalty Builders Finite State Machine model can include multiple variables without making these trade-offs.  The result is a richer, more accurate picture of customer behavior.