Application of Customer Relationship Management and Data Mining in Predicting Customer Purchase Behavior in Medium and Large Supermarkets in Kenya
View/ Open
Date
2015-07Author
Nyoike, Catherine W
Type
ThesisLanguage
enMetadata
Show full item recordAbstract
The study topic was inspired by the growth in research dynamics for futurist organizations.
Conventional research has concentrated on traditional methods that apply primary research
and secondary research. One of the industries that has experienced rapid growth is the retail
industry, so much that global brands like South Africa’s Shoprite have had a hard time
penetrating the Kenyan market. Supermarkets are great custodians of big data which when
mined can provide meaningful insights on basic organizational process as well as consumer
behavior. The study thus sought to ascertain the extent of application of customer
relationship management and data mining in predicting consumer behavior in medium and
large supermarkets in Kenya. Random sampling was used for the study which was carried out
amongst 30 supermarket staff in managerial cadre. Data was collected using selfadministered
questionnaires and face to face interviews which were also done to get optimal
responses from the respondents and analyzed using descriptive statistical measures like
frequency distributions and percentages. The study established that while CRM is applied in
the retail industry to some extent, basic data mining techniques are still popular for mining
insights and are generally not specialized for data mining functions. It thus emerged that
Excel is the most widely used software at 92% with usage being mainly for measurement of
sales and customer purchase frequency. On sources of customer data and the respective
applications, it was established that loyalty cards would ideally provide the most
comprehensive customer data. In terms of usage, it emerged that the current retail
technologies are used mainly to capture sales at 100%, measure customer purchase frequency
at 93% and measurement of stock movement at 90%. However, various variables were not
captured due to use of manual systems. Based on the findings various recommendations were
drawn, key amongst them the need to invest in advanced data mining systems and
diversification of data types that can be captured by such technologies including and not
limited to customer demographics, customer complaint handling mechanisms and tracking of
lapsed customers. These shall be steered further by use of systems that capture data
comprehensively, at best this can be captured by loyalty cards whose usage can be increased
through campaign programs by the various retailers. There is still a long way to go for the
industry to grow and maximize its data mining potential. Consequent research may look to
establish reasons for the low uptake of data mining technologies in the industry given its
known potential for a higher return on investment (ROI).
Publisher
University of Nairobi