Problem statement

In today’s competitive market, understanding and leveraging Customer Lifetime Value (CLV) is crucial for businesses aiming to maximize their return on investment (ROI). As we all know Return on Investment is the foundation upon which every business is build.

CLV is a predictive metric that estimates the total revenue a business can expect from a single customer account throughout their relationship. By focusing on high CLV customers, businesses can optimize their marketing strategies, particularly in customer acquisition.

One powerful strategy to acquire high CLV customers is using lookalike audiences on social media platform such as Instagram. We will show case how one of our client has leverage CLV, specifically calculated using the Beta-Geometric Negative Binomial Distribution (BG-NBD) model, to perform targeted customer acquisition through Instagram’s lookalike audience feature.

Concept and definitions

What is CLV ?

Customer Lifetime Value (CLV) is a crucial metric for businesses, representing the total revenue a customer is expected to generate throughout their relationship with the company.

What are Lookalike Audiences?

Lookalike audiences are a Meta feature that allows businesses to reach new people who are likely to be interested in their business because they share similar characteristics with their existing best customers. This feature uses data such as demographics, interests, and behaviors to create a profile of your ideal customer (ideal here means the one that brings high revenue 🤑).

• Precision targeting: Reach potential customers who are similar to your high CLV customers.

• Increased engagement: Lookalike audiences are more likely to engage with your content, leading to higher conversion rates.

• Cost-effective: By targeting users who are more likely to convert, businesses can reduce advertising spend and improve ROI.

The purchase process is parametrized as follow:

• Recency (r): Time elapsed since the customer’s last purchase. Recent purchases indicate higher engagement and potential for future transactions.

• Frequency (α): The average number of purchases a customer makes within a given timeframe. Customers with frequent purchases exhibit higher loyalty and likely have a higher CLV.

The Beta-Geometric Negative Binomial Distribution model uses two probability distributions to model customer ltv:

1. Gamma distribution to model purchase behaviour: Each customer has an unobserved “propensity to purchase” that follows a Gamma distribution with shape parameter $r$ and scale parameter $\alpha$:

2. Beta distribution to model dropout process: After each purchase, a customer may become inactive with a probability that follows a Beta distribution with shape parameters a and b.

In the end BG/NBD model uses four parameters to describe the rate at which customers make purchases and the rate at which they drop out.

Conclusions and discussions

Let’s remember that “All models are wrong, but some are useful.” Predictive Lifetime Value (LTV) models are especially valuable for understanding how actions taken by teams, such as acquiring and engaging customers, influence important outcomes like revenue and customer retention. These models help map the relationship between what teams do and the value they generate, making them useful despite their inherent imperfections.