Ramgopal Prajapat:

Learnings and Views

CLTV definition and CLTV for Customer Acquisition

By: Ram on Jun 12, 2022

In the previous blog, we discussed about the applications of Customer Lifetime Value (CLTV). Now, we will get into more details to define CLTV and the approaches for measuring and managing CLTV for customer acquisition.

For Customer Lifetime Value Measurement, we need to define

  • Who is a customer?
  • What is lifetime for a customer?
  • What is the “Value”

 

Customer

For an ecommerce business, there can be multiple ways define period as its customer and few of them are:

  1. A visitor to the website or who has installed the mobile application
  2. When a person registers on the App or Web
  3. When a customer makes its first purchase

Of courses, there are pros and cons of defining a person as customer using any of the above methods. Considering the commerce angle, we tend to consider the 3rd option for define customers for the CLTV measurement. As soon as a person makes his/her first purchase, s/he is a customer and the CLTV will be predicted (so it is a future looking metric) for him/her

 

Lifetime

Before, we get into the detailed modelling, we need to define the lifetime period for a customer. Though, theoretically, it is lifetime of a customer but practically we need to define a fix period for the value computation or measurement.

Typically, 3 or 5 years is considered as a lifetime for a customer with an organization.  In the current scenario, our focus is around CLTV for “customer acquisition” and what factors drives the customer value.  Find the list of controllable factors or features that helps us in finding the high value customers at the time of acquisition.  So, one view is to consider the value added by the customers in the first 6 months as post 6 months the engagement and value could be linked to various other factors e.g.  engagement program of the ecommerce organization. Additionally, if we need to consider the customers acquired in the latest periods, we may not have the longer period data for CLTV measurements.

So, we may consider 6 months window to measure value of customers. The construct to define the target or outcome value is depicted below.

 

Timeline

Description automatically generated

 

Also, if we look at the data, the customers who were identified as high value based on the first 6 months remained high value for next 21 months from the first purchase.  So, the first 6 month is good enough period to segment the customer into high and low value groups.

 

Chart, line chart

Description automatically generated

For a customer who have been around for over 6 months, we can consider 2/3 Years for CLTV measurement. Again, considering dynamic business construct and evolving customer expectations, probably shorter period 2/3 years is advisable.  

Value

Depending on the focus and scenario, the value can be function of various factors. For an ecommerce business the value is function of following

  • Order Amount
  • % Commission (for marketplaces) or Sales Margin (Inventory Led Model)
  • Delivery Charges Revenue
  • Cancellation and Return Costs
  • Promotion Costs
  • Cash on Delivery (COD) or Payment Gateway Charges
  • Logistic Costs

 

For the acquisition scenario

CLTV is computed as

CLTVn=Total revenuen-Total costn

where CLTVn is CLTV measured for each customer (1,2, 3…n)

Revenue

  • # Of Orders * Average Order Value * % Margin/Commission + Delivery Charges Income

 

Cost

  • Cancellation or Return Costs + Logistic Cost + Payment Gateway/COD Cost

 

Marketing and other fixed costs are sometime difficult to attributes to a customer.  But, if possible, we should consider these costs especially the acquisition marketing cost so that relative value and cost of acquisition channel or platform can be considered.

 

Acquisition CLTV Computation in summary

  • Consider first 6 months post first transaction as a period
  • Only customers who place their first orders are eligible for CLTV computation
  • Value considers all the cost and income components that can be calculated and attributed at a customer level

Now, we will discuss on the methods to predict pr forecast CLTV at a customer level

  1. Customer Level CLTV Model using Regression based Methods
  2. Customer Value is segmented in 3 buckets and then a customer value classification Model is built
  3. Model Individual Components
    1. Predict # of Orders for a customer in next 6 months
    2. Estimate or Forecast Average Order Value
    3. Blended Commission Rate or Margin is taken
    4. Estimate cost

Leave a comment