For data analyst freelance services contact me [email protected]
In data driven companies, there is a lot of information about their customers stored in the database. However, it is recommended to add fields with holistic (360) information that will be used for customer analysis. One method of creating such fields is through a customer scoring model. This model allows for the consolidation of several variables that describe the customer into a single score variable.
In the next post, I will describe this model.
Every company keeps extensive information on its customers according to its business model. Customer information can include: frequency of purchases for each customer, annual purchase amount, number of times the customer used the product, whether the customer returned the loans provided to them, and so on.
Each field in such a database represents one aspect of the customer, but customers are complex beings and in order to assess their value to the company, it is necessary to look at multiple aspects together. With such holistic analysis, the company can perform a complete customer analysis and know which customers it is worth keeping and trying to recruit similar customers, and which customers can reduce preservation efforts.
For example, if a particular customer purchases one high-priced product, they may not necessarily be as valuable to the company as a regular customer who purchases lower-priced items, or customers who take out high amount loans but have difficulty returning them may not be more valuable than customers who take out lower amount loans but repay them regularly.
The score can be composed of a large number of variables. Some of the variables may have very little weight and have little effect on the score, while others may receive negative weight (such as in the case of customers who do not repay their loan).
Holistic score for customers allows to aggregate the customers and have a comparison between segments. Such analysis can dramatically change the way the company relates to certain segments. For example, if it is discovered that customers in New York have a higher score than customers in Chicago, the company may prefer to try more New York customers and perhaps even close branches in Chicago.
Variable selection and weight determination
This is the most important stage in the process and it involves business thinking. The data analyst needs to interview the business stakeholders and understand which variables are most important for the business model.
Before this stage, it is impossible to proceed and the discussions of determining the most important variables can also raise important strategic questions that the company needs to answer.
The output of this stage should be a list of the important variables and the weight in percentages of each variable (the weight percentages should add up to 100%).
Normalize variables
After selecting the variables using business factors, the data analyst needs to ensure that all the variables speak the same language. This process is called normalization, and there are several statistical methods to do it.
The best way, in my opinion, to normalize variables is by transforming them into standardized scores. Another method of normalization is by dividing all variables by the maximum value of the variable, but this method may be influenced by extreme values.
Creating the new customer score variable
After normalizing the variables, the data analyst can create the new score variable using a simple linear transformation. Each normalized variable is multiplied by its assigned weight, and the score for each customers is the sum of all these products.
Sanity Check for the new customer score variable To summarise
At this stage, when everything is ready, the data analyst needs to perform a salinity check on the new score. In the test, it is necessary to ensure that the customers with the highest score are indeed the most important customers for the company and that the customers with the lowest score are indeed not good customers. Usually, it is found that some adjustments to the weights we determined at the beginning are needed.
To summarise
In order to perform customer analysis and gain a holistic view of the company’s customers, the data analyst can create a score for each customer composed of several variables, each with a different weight according to the company’s business model. Using the score, it is possible to identify the most important customers for the company and retain them. Aggregation of the score also enables ranking between different segments and changing the company’s strategy for those segments.
This article was written by Yuval Marnin.
If you have a need to hire a freelancer data analyst you may contact me at: [email protected]
You may also hire me through upwork platform on that link:
https://www.upwork.com/freelancers/~018940225ce48244f0\
More articles on Yam Analytics blog
The advantage of hiring a freelance data analyst.
What does a data analyst is doing and how it can help your company.
