← blog

Purchase Propensity Model, An Angel in Disguise

Businesses reportedly spend 7-8% of their revenue on marketing and advertising. In 2019, the global digital advertising market value stood at $325 billion. As a result of Covid-19, businesses were cautious about their advertising efforts in 2020. But with things getting back to normal, the number has been reported to cross the $389 billion mark in 2021.  

Businesses spend hundreds, if not thousands, of dollars on customer retention. However, they struggle retaining their already-existing customers. That’s because they aren’t familiar with their customer’s purchasing pattern and behavior.

As a result, they waste millions of dollars each year reaching out to the wrong audience. This results in an extremely poor ROI.

But what if you can now identify the customers with a higher probability of purchase?

Introducing the ground-breaking addition to Predictable Media Customer Data Platform:

“Purchase Propensity Model”

Businesses can tap into the power of this model to identify the customers with a high probability of purchase.

This article shines a light on the Purchase Propensity Model. Also, we have incorporated a real-life example to explain how it works.

Let’s dive straight into it.

What is the Purchase Propensity Model?

A successful purchase isn't the end. Your focus should be on retaining your already-existing customers – as many of them as possible.

Honestly, you can’t have a 100% customer retention rate. That’s literally not possible. Different individuals have different tastes and behavior. While some may really like your product(s) or service(s), some may not.

Let’s say one of your customers purchases your product(s) or service(s) every month. On the other hand, another one of your customers buys from you at the end of each year – maybe Christmas.

So, if you are planning to roll out an offer in July, who’d you target?

My guess – it’s the customer who purchases your product(s) or service(s) frequently.

Reaching out to every single one of your customers is a dumb move. Instead, the Purchase Propensity Model allows you to identify the customers who are most likely to purchase from you in the next few days.

“Purchase Propensity Model is a widely recognized statistical approach used by businesses to estimate the likelihood of their target audience taking the desired action.”

Businesses tap into the power of this model to predict their target audience's behavior. This model is based on the Gradient Tree Boosting algorithm. It combines predictions from numerous decision trees to generate the final predictions. Here, every successive decision tree is created after taking into consideration the errors of the previous trees.

This model used the RFM (Recency, Frequency, andAmount) features at 1, 3, 6, and 12 months and focuses on discovering non-linear relationships.

Predictable Media experts have developed two functionalities to help you get the most out of it:

·     PurchaseLikelihood for the next _ days

·     Success Probability

Let’s look at how this model works practically.

Purchase Propensity Practical Approach

Let's hop onto our Predictable Media platform. Once you have logged in, select "Audience Creator."

Here, we have selected our “Shopify E-commerce" data source from a glasses company. Now, let's unlock the full force of the Purchase Propensity model.

As you can see, here’s what it looks like.

Purchase propensity model in Predictable Media Software.
Purchase propensity model display in Predictable Media Platform.

So, first, let's say we want to target the customers who are most likely to buy from us within the next 15 days.

So, we’ll set the “Purchase Likelihood for the next”to 15 days.

And let’s set our Success Probability to 70-100%. This will help us target the customers with a 70-100% probability of purchase within the next 15 days.

And here we go:

Data Science models in Predictable Media Platform.
Data Science models in Predictable Media Platform.

Our algorithm has narrowed it down to 42 customers.Now, let’s say you want to target the customers who are most likely to purchase a specific product.

So, all you got to do is head down and select the product.

Here’s a list of our products:

List of Products of Some Glasess Shop in audience creator module.

So, let’s select all “ClubMaster” products

Audience creator module in Predictable Media Platform.
Select all Rayban ClubMaster products of Some Glasess Shop in audience creator module.

That’s twenty five customers we have.

And you can sync or export this audience to reach out to these customers.

Or you may even use other filters to further narrow down your audience.


Purchase Propensity Model is a widely popular approach. Predictable Media allows you to unlock its full force within a few clicks.

It's simple, and it's brilliant. Based on this, you can create high-impact campaigns and level up your marketing efforts.

So, what are you waiting for?

Hop onto Predictable Media Customer Data Platform today.  

Related articles