When performing this analysis, you need to take into account a lot of information about customers. The results of the study can be used to develop targeted promotions. Do not forget that the results of segmentation need to be verified in real life, because differentiation into clusters can vary greatly from month to month. In addition, such segmentation can be used to analyze surveys. At the same time, text information is not so easy to convert into numerical indices, especially if you need to process a large number of respondents. Therefore, you should ask questions like: "Rate the importance/quality/size of ... from 1 to 5."
Example. Let's say you are the vk database owner/employee of a credit and financial institution. Clients should be divided into groups depending on which banking offers they use. At the same time, for each product, several questions on the importance of selection factors should be developed. The respondent will have to rate each characteristic on a scale from 1 to 5. Let's look at an example of the result of such segmentation:
Debit card holders:
economical – the highest score was given to the factor “cost of annual maintenance”;
use the product for transfers - pay attention to the size of the commission for transfers to cards of other banks;
conformists – value factors such as “brand reputation” and “reviews” more than “cost of service”.
Legal entities regularly performing cash transactions:
small entrepreneurs – for them, factors such as “the cost of opening an account”, “the convenience of connecting and using the bank’s Internet services”, “favorable service rates” are important;
companies with large tranches focus on cash transaction limits, as well as on the reliability and reputation of the banking organization.
Target audience segmentation using association rule analysis
Association rule analysis (market basket analysis) is used in the process of finding stable combinations of goods in purchases. Various algorithms are used for this. AIS is the very first of them - it was created back in 1993. To perform the analysis, you will need data on purchases. In this case, each transaction must have its own identifier (usually the receipt number is taken) and the items included in it.
Target audience segmentation using association rule analysis
Companies that are not in the FMCG segment can replace the receipt number with a unique buyer ID. This will allow you to identify systematic behavior patterns of customers based on their purchase history. Based on this information, a list of recommendations will be generated. For example, 4,000 customers were identified on Aviasales, and 1,400 on Booking. At the same time, 700 people were registered who made transactions on both platforms at once. The total customer base is 6,500.
Based on this information, two indicators are calculated: confidence and support of the rule. The latter is the proportion of customers who made purchases on both platforms from the total number of transactions. In our example, this is 10.8%. Confidence (connection strength) is the proportion of customers who made purchases on both platforms from the number of transactions of each of them separately. Obviously, confidence has two values. In our example, for Booking it is 50%, and for Aviasales - 17.5%. Thus, users are more likely to make a purchase on Booking, and only then on Aviasales.
How will this information help in marketing? It is best to create a promotion that will primarily advertise Booking, because after making a purchase on this platform, customers usually go to Aviasales. In addition, you can adjust the automatic mailing as follows: after a transaction appears on Booking, the corresponding buyer will be sent a promo code for a purchase in Aviasales with a discount for a limited time. An excellent way to increase income will be to create a combo set. In this case, when buying it, a person can count on increased cashback.
Download a useful document on the topic:
Checklist: How to Achieve Your Goals in Negotiations with Clien