Introduction to Customer Analytics


In our last article we learnt about six important classifications of Analytics. In this article we will look into details about Customer Analytics and the various techniques involved in the extensive process.

What gets measured, gets managed - Peter Drucker, Analytics Expert

Let's understand more about Customer Analytics with this video


In this video we defined Customer Analytics as the Analytics based on customer behavior and transactions. Shopping patterns are analysed to give the best possible deals, to attract and retain a customer.


Valve software and it's lovable CEO, and co-founder, Gabe Newell, is much respected for his ability to manage his billion dollar company without a single manager. Gabe (Often referred as to Gaben) is considered to be one of the most responsive CEO’s to Fans and Clients. Over the course of 25 years of Valve Software, he has crunched big numbers in both software and hardware aspects. Having had his experience in Microsoft's OS dev team, he knows the industry in and out and how important a customer can be. In an interview he stated,

“There is only one boss, the Customer. He can fire every single person in a company, from the clerk to the Chairman, just by spending somewhere else..”

With the importance of a customer, comes the need to attract and retain more. Customer Analytics plays an important role when it comes to strategizing these aspects.


Now let's discuss the few prominent steps of Analytics here:


1.Understand the Customer -

The company needs to evaluate and draw out the needs of the customers. The Customers transactions are analysed and a profitability report is drawn. This report helps the company determine whom to focus on. Almost every company uses this extensively, this single handedly is the greatest tool to solve companies problems. This involves concepts of Percentile and profit/visit. For Example,

Your search history on an E-commerce site (say Amazon) is tracked and is tracked for what you've purchased. Often when you buy a phone, the primitive advertisements would be about Screen protectors/Back Cases. These are attempts to understand your needs and provide convenience.


2. Classification -

Customer behavior is analysed in three different ways, and are rated by a ML algorithm. The basis of rating is an RFM algorithm. RFM stands for Recency, Frequency and Monetary basis. Customers are segmented from each category and given an average score to every profile and ranked. On basis of

ranks, the company decides what the best strategy to approach a customer would be. For Example, A company might devise various “Expenditure” limits (Like Cisco did to boost sales in 2008) and categorizes each customer based on how much they have spent. Based on those ranges, products of budget, mid range or premium products to various customers.


3. Customer Loyalty -

Analysis on why a customer chose the product/service is very crucial. Based on customer inputs, a company can either work more on a particular channel or strengthen a weak one. The easiest method to do this is a NPSA (Net Promoter Score Analysis). This is based on two standard questions, “How likely would you recommend company/product to your friends/family/associates?”, and, “How did you get to know about our product/service?” For Example, with each update of Microsoft Windows 10, you will receive an email from Microsoft asking you how likely would you recommend Windows 10 to your friends/colleagues. Microsoft uses this data to figure out which update was helpful to the users and which update was not, helping them provide the best experience to their users.


4. Potential Profits -

Based on the above inputs and the customer behavior, his/her net potential for profit is calculated and a prediction of how profitable a customer would be to a company is evaluated. This is called a “Lifetime Value”. This kind is mainly used for advertisement purposes only. For example, e-commerce sites advertise similar products of what you've purchased to promote sales. The last step is more of a trial and error format without an extensive use of it.

With these methods, great detail is put into rating each profile to create targeted campaigns to expect high ROI’s, in a short amount of time.


Thus we conclude this article here, in next article we will learn more about techniques used in Customer Analytics


If you liked this article then "Subscribe" to our YouTube Channel for more updates:



0 views

© Markytics