Every business, every industry faces attrition. Customers leave for a variety of reasons which are easily recognizable. For example, customers stop purchasing disposable diapers when their children no longer require a diaper, automobile owners may no longer patronize the car dealership when the warranty is no longer in effect. A competitor's new products might cause unforeseen churn. While there are many logical, rational and understood reasons that a customer leaves a company it is the silent attrition which can be the most devastating in this economy.
What is silent attrition?
Silent attrition is a situation in which a customer or client base discontinues patronizing a business without any explanation. According to Andrea J. Ayers, president of customer management at Convergys, “Silent attrition varies by industry, but in general, companies are losing about 12% of their customers this way, with the defectors poisoning the well among potential new customers. Silent attrition, particularly in this economy, can spell the difference between a company’s success and its failure.”
Typical attrition models
Typically, attrition models use historical behaviour of customer to segment them based on their propensity to attrite. In other words, they are a statistical model designed to predict attrition within certain segments. A predictive model can be built using the observable customer attributes derived from CRM data which defines both loyal and attritors and then applies those attributes to other customers.
Using this predictive model the customers can then be segmented in to high risk – likely to attrite and low risk – likely to remain loyal groups. Retention campaigns can then be applied to the high risk groups. While a tried and true methodology, predictive models do little to stem the tide of silent attrition.
Advancing to stem silent attrition
What is missing from the typical attrition models is what is known as customer intelligence or how the customer’s interactions with the company and the company’s support of those interactions are perceived. This type of unstructured data can be amalgamated into the CRM solution and then classified or segmented for structure.
When the data is collected, either through outgoing sales solicitation or incoming call center activity the dialogue is classified into a tabular format. The categorization of the dialogue is through text mining for significant or strategic words. By combining structured data with the unstructured data this table represents a predictive model that can be constructed which can assist in stemming silent attrition.
Using the amalgamated data customers can be scored on their propensity to attrite. Scoring would be based on the model developed using both the CRM structured data and the customer interaction or unstructured data. The combination of the two data sources in the scoring model provides improved predictive qualities.
When scored, retention programs can then be targeted toward those customers most likely to attrite who have the potential to be profitable.
It is an established axiom that there are greater costs associated with the acquisition of a new customer than with the retention of an existing customer. With the current state of the economy, retention becomes an even more important aspect of every company’s business plan. By using a predictive model which integrates structured CRM software data and unstructured tabulated data, the company is more likely to have success with retention programming than a company which relies solely on historically predictive data.
Comments