Market basket analysis
What is market basket analysis?
A customer walking into a retail sector like a supermarket is likely to make purchases that are not previously planned. Often, there are impulsive purchases triggered by the price and position of other products within a physical store. At such stores, a customer may decide to purchase a product after entering the store. There are patterns of purchases that can be assessed on the basis of products bought together. Market basket analysis is a crucial technique used by retailers to increase sales by evaluating purchase patterns. It is a data mining technique that identifies associations between products purchased together.
By looking at the purchase pattern of the customers, businesses can increase the profit margins through promotions, recommendations and cross-selling. Certain products within a store may have the potential to attract other purchases. These keystone products can influence the overall sale of a business. One main factor that affects a company's profitability is the number of people who view the products. Keystone products can act as a marketing strategy to increase the sale of other products. Identifying these products is important to a business, as the availability of these products within the store influences the number of customers walking into the store in a day. Market basket analysis is extensively used in online stores to recommend other products. It helps in developing product placement and sales strategies.
Figure 5: Recommendations of market basket analysis [1] |
Why is it used?
Data for market basket analysis can be obtained from the transaction level. Depending on the aim of the analysis, the algorithm can be fed with overall transactional data, segment-specific or customer-specific transactional information. These data are often easier to obtain, as transactional data are rarely ignored. In cases where data is not recorded, a data collection framework can be built.
Based on the theory that a customer is more likely to purchase some products when certain other products are bought, market basket analysis helps businesses cross-promote items to the customers. Simple market basket models built based on support, confidence and lift. Support refers to the transaction that contains all items in a dataset. Confidence is the probability that a transaction containing some items also includes certain other items. Lift refers to the likelihood of all these items occurring as if there are no associations between them. Thresholds are set for support and confidence based on the domain. Business sectors from hotels to supermarkets to online stores extensively use market basket algorithms to cross-sell items. Businesses can use this technique to change the store layout to group some items, design catalogue designs, assess top items, customise marketing strategies, and much more. For online purchases, the algorithm helps in recommending products when other products are purchased. These can be used to build loyalty incentives and thereby reduce customer churns. In the banking sector, it can be used to reduce fraudulent transactions by analysing card purchases. In the health sector, it is used to reveal biological associations between active components.
Contact us to know more:
statmlconsulting@gmail.com
Comments
Post a Comment