Pre-collection precautions for data accuracy
A study specific to Mobile Industry


By
N.M.Shanthi
Ph.D. Research Scholar
Department of Management Studies, Anna University
Chennai
E-mail :
shan_venky@yahoo.com
 


".but this is not what I got from feedback!" yells a startled Chief Customer care Officer (CCO), of a cellco giant looking at his quarterly churn figures. Many CCOs, especially in the service industry are experiencing this situation quite frequently. They are able to sense the enormity of its impact on the overall operations of the company and hence are sweating out on focused exercises to identify probable reasons for variation in the actual and projected figures on customer defection, especially when it creeps much beyond the acceptable range (when it crosses the higher limit. If it is below the lower limit, they are obviously happy).

With the present day requirement to necessarily hold back the customers for sustainability, companies frame tailor made strategies on a timeline basis, through vibrant and customized CRM solutions available with them. They have realized that the cost to retain a customer is not only much lesser than the cost incurred to get him but also is a high yielding investment for the company's sustainability. "4 loyal customers of today are 400-tomorrow", is the continuously chanted mantra by walls, tables and chairs of the customer support departments in today's customer-centric organisations.

Conducting constant and well-directed feedback surveys to measure the expectation and satisfaction levels of the customers, has become an essential part of the customer support activity. Each company has its own way of conducting feedback surveys. Some follow the standard methodology, while some design their own methodology, depending on the nature of the product and the size and composition of the customers. The customer research groups (in-house or external) conduct regular surveys to get the feedback of the customers, so that measures can be taken to minimize the gap between the expectation and satisfaction levels of the customers, in order to retain them. CRM strategies get reframed or modified based on the interpretation of customer expectations as given by the customer feedback analysis.

Despite all the focus on customer retention, customers defect. A defection rate of 5-8% is generally acceptable to most of the companies, especially companies with very large customer base due to unpredictable behavioral patterns. This rate is bound to increase or decrease based on the results of the customer feedback surveys. With a 5% acceptable variation between the actual and expected rate of defection, companies predict growth prospects for the future, which form the base for framing enhanced at the same time target-focused marketing and customer support activities.

Things are not always as green as you want them to be. There are instances when CCOs are left to face a situation when the variation between the actual and predicted rate of customer defection goes much beyond the acceptable higher limit. The immediate impact of the disaster falls on the customer care executives, who are the closest to the customers. A process of corrective measures that follows would range from upgrading the skills of the executives to reshuffling the groups or even sacking a few for reasons of incompetence.

While reasons for increased defection could be due to many external factors, there is an important factor, which could have caused a huge variation between the actual and predicted figure and that is hidden flaws in the survey methodology. Any methodological imperfections at the grass root level, would lead to disastrous calculations. No company can claim to have adopted a perfect survey methodology, but the imperfections, if any could be corrected through statistical adjustments done post-data collection. In case there are any unintended flaws in the data collection methodology adopted, the actual figures arrived at may be true but the variations could be due to misleading predicted figures.  There are certain pre-collection precautions, which need to be taken to assure better accuracy of the data collected.

  • Customizing questionnaire:  The first activity any company would do before implementing CRM strategies is customer categorization. Customers are categorized into identical groups, based on their demographic and geographic and product preference and usage characteristic. Each group needs to be surveyed separately using questions based on unique characteristic they posses. These customized questionnaires provide better insight into customer expectations and assures better accuracy.
     
  • Changing the questionnaire pattern: The question pattern needs to be changed for each and every survey, to make it interesting to the respondents. Regular questions make it boring and would fail to extract the required information.
     
  • Deleting obsolete questions: One of the major flaws in the questionnaire used for the survey is usage of obsolete questions. A question becomes obsolete when all respondents express high level of satisfaction on that. These questions need to be identified periodically and deleted from the questionnaire, to avoid misleading inferences.
     
  • Method of data collection: Many companies send feedback forms to the customers and collect them either by mails or e-mails. The response to these kinds of feedback mailers is generally 10%. So it is recommended that direct feedback collection is done to increase the response rate. An increased response rate leads to increased accuracy in the data collected.
     
  • Short questionnaire: Studies reveal that short and focused questionnaire promises better accuracy than long ones, because long ones test the patience of the customers, and by the time they reach the last part, which could be the most target focused part of the questionnaire, they partly of fully lose their interest to answer.
     
  • Rating scales and ranking: Care should be taken in selecting the rating scales and ranking list. An even rating scale like a four point scale can be used without a neutrality point, in order to force the respondent confirm to a particular positive or negative rating, since neutrality would not give adequate information about the customer's preferences. As far as ranking is concerned, number of variables taken to rank should be minimal, so that the respondent finds it easy to rank the parameters based on his perceptions.
     
  • Reframing questions: Using emotionally powerful words like "Fantastic" for "Very satisfied" and "Disappointing" for "Not Satisfied" would make the questionnaire better focused  and interesting. This would also lead to better understanding of the customer expectation and satisfaction levels.
     
  • "Satisfied" are not necessarily loyal: "Satisfied? Who cares?" is the general attitude of all customer support groups. But, they should realize that customers who profess their satisfaction switch to rival products in no time. So keeping a constant monitor of the feedback from the so called "Satisfied" groups is as important, because satisfied need not necessarily is loyal.
     
  • Less time gap: The time gap between the data collection and analysis should be very less. In behavioral studies the data cannot bet stability for a long period. If the data lies in the shelves for more than 3 months, it is declared dead. Any analysis based on that would only lead to misleading predictions.

All the pre-collection precautions stated above should necessarily be taken to assure optimum accuracy in the data collected. If a CCO says that he is been taking all the precautionary measures but still sees variations, he should definitely check the operational and functional efficiency of his company, by performing a SWOT analysis. This would enable him to identify the deficiencies, and take corrective measures to reduce customer defection.

 


N.M.Shanthi
Ph.D. Research Scholar
Department of Management Studies, Anna University
Chennai
E-mail :
shan_venky@yahoo.com
 

Source : E-mail September 29, 2004

 

B A C K

 

Important Note :
Site Best Viewed in Internet
Explorer in 1024x768 pixels
Browser text size: Medium