1 Paper 1267-2017 Behavioural spend modelling of cheque card data using SAS ® Text Miner Amelia Van Schalkwyk, Frans Kanfer, and Sollie Millard, University of Pretoria. ABSTRACT Understanding customer behavior profiles is of great value to companies. Customer behavior is influenced by a multitude of elements-some are capricious, presumably resulting from environmental, economic, and other factors, while others are more fundamentally aligned with value and belief systems. In this paper, we use unstructured textual cheque card data to model and estimate latent spending behavioral profiles of banking customers. These models give insight into unobserved spending habits and patterns. SAS® Text Miner is used in an atypical manner to determine the buying segments of customers and the latent buying profile using a clustering approach. Businesses benefit in the way the behavioral spend model is used. The model can be used for market segmentation, where each cluster is seen as a target marketing segment, leads optimization, or product offering where products are specifically compiled to align to each customer's requirements. It can also be used to predict future spend or to align customer needs with business offerings, supported by signing customers onto loyalty programs. This unique method of determining the spend behavior of customers makes it ideal for companies driving retention and loyalty in their customers. INTRODUCTION Businesses are driven by their ever changing goals and aspirations which may differ from large corporations to sole traders, however they all have one very significant commonality, aiming to sell their products and services. Whether companies are selling groceries, bank accounts or fitness classes, their products are their underlying structure. For this reason it is of paramount importance to each company's continuous success to sell their products efficiently. In the fluctuating uncertainty that arises from environmental, political and economic change, it is essential for companies to understand the needs, constraints and aspirations of their customers. To do this it is necessary to delve deeper into the behaviour of customers, this type of modelling can be known as a form of behavioural modelling that stems from fields such as social and behavioural sciences. The term behaviour is fairly broad, for example it can range from how often a customer travels to what time of day they spend their money. Consumer behaviour falls within the fields of social and behavioural sciences as it is a component of social behaviour. Krishna, [3], defines consumer behaviour as “A discipline that encompasses all processes involved in acquiring, using and disposing of products, services and ideas. As a discipline it is also concerned with the results of such acquisitions, uses and disposals”. This paper focuses primarily on the acquisition side of consumer behaviour, more specifically the purchasing/buying/spending behaviour of the customers and not the psychological reasons behind acquisition. Behavioural spend modelling is one of the many possible consumer behaviours that can be modelled, it is the process of detecting and understanding patterns in the spend data of customers. These identified patterns assist in determining behavioural segments and spending profiles. Interpreting the model and recognising whether customers spend more on groceries, holidays or restaurants, etc., in other words whether they spend on luxury items or necessities, assists in understanding the customers lifestyle and unique needs. SAS® text miner is used in an atypical manner to determine the spending segments and latent spending behaviour of customers using cheque card data. The latent spending behaviours of a customer are in turn used to determine the customers spending profiles. The cheque card data is transformed in such a way that the result consists of a combination of spending terms making up a customer’s spending document. This assists SAS® text miner in capturing all the spend in each customer document as behaviours. The results provided in this paper suggest the use of SAS® text miner as a tool for modelling behavioural spend is a successful one, the customer selected at random and analysed produces results indicating a lavish lifestyle, this could identify a certain set of products uniquely suited to this customers lifestyle. This