Blog

3 Machine Learning Models That Will Help Boost Marketing Revenue

See examples of three types of machine learning models marketers can use in Acquia Marketing Cloud to understand the true value of every customer.

As marketing grows increasingly data-driven, businesses are turning to artificial intelligence and machine learning (ML) to help identify patterns and draw insights from big data sets that will help them create more effective campaigns. The use of machine learning models to help marketers anticipate future customer preferences and more accurately offer a relevant action is predictive marketing. Mastering predictive marketing means more personalized and relevant experiences for customers and stronger, more valuable relationships for brands.

Acquia Marketing Cloud employs a robust machine learning framework that allows marketers to leverage both pre-built and configurable machine learning models, so they can customize their analysis and predictions to be in-tune with their unique business goals. Our recent enhancements to the Acquia Digital Experience Platform introduced additional machine learning models to serve a wider range of use cases and provided the ability to use fuzzy clustering to group customers into machine learning segments for more accurate segmentation. Fuzzy clustering is a technique that expands the size of ML segments by allowing individuals to be grouped into more than one segment at a time.

To date, our marketing cloud currently offers 14 different machine learning models out of the box, so teams can align their segmentation and personalization tactics to greater business goals. Here are three powerful machine learning capabilities that can help marketers grow revenue, improve customer relationships and grow retention rates. 

1. Likelihood to Pay Full Price

Discounting seems like a no-brainer for brands to boost slow business. Nothing gets a customer’s attention faster than a big flashing “SALE” sign or a “90% off! Limited time only!” flyer. However, when every day there’s a new special offer, the offer starts to feel not so special. In fact, discounting can actually hurt business by lowering your margins and de-valuing your brand image. Instead of offering blanket discounts to all customers, machine learning can help brands better understand their true worth. 

The Likelihood to Pay Full Price (LTPFP) model in Acquia Customer Data Platform (CDP) predicts the degree to which a customer is likely to purchase a product without a discount. By identifying these cases where additional discounts are unnecessary, brands can both increase their revenue by getting more full-price conversions and reduce the time and resources spent sending irrelevant discount emails and offers to this audience.

Some ways a brand could use the Likelihood to Pay Full Price model to optimize their campaign efforts include excluding customers in the LTPFP segment from coupon promotion blasts or giving lower discounts to customers that are likely to pay full price while giving more aggressive ones to customers that are not. This prevents brands from shooting themselves in the foot and hurting their profit margins. Check out the Likelihood to Pay Full Price model in action below:   

2. Predictive Lifetime Value

Ask anyone if they’d rather have hundreds of fickle, temporary friends or a few trusted, lifetime  confidants, and it’s likely most will go with the latter option. The same can be said for relationships between brands and their customers. Customer retention and strong, long-term relationships are becoming increasingly important to every company’s success as more and more channels, content types and services pop up every day to try to pull customers’ attention. Studies have shown it costs businesses five times more to acquire a new customer than to retain an existing one. So, it's important that today’s marketers understand who their ride-or-die brand loyalists are and invest in a good customer retention strategy.

The Predictive Lifetime Value (PLV) model predicts the revenue or margin of a customer over the course of the next 12 months. The machine learning capabilities calculate each customer’s predictive lifetime value by looking at their actions across different channels from in-store purchases to email engagement to web-based behaviors like session duration and cart abandonment. 

By identifying the most engaged, highest value customers, marketing can focus their efforts on nurturing this existing core audience to buy more, buy more often and recommend your company to others. For instance, you may want to entice brand loyalists with exclusive offers and events or reward them for their business with a special gift designed to surprise and delight them. Beyond access to perks and special content, brands could also reap amazing returns by turning customers in their highest lifetime value segments into active, enthusiastic brand advocates and influencers. You can target these customers with opportunities to submit user generated content or leave reviews in exchange for earning points.

See how the Predictive Lifetime Value machine learning works in our demonstration:

3. Next Best Channel 

From the MySpace craze of the early-aughts to the massive shift of Facebook from a space for college kids to one for middle-aged parents to Gen Z’s trending TikTok dance crazes, each digital channel attracts different types of content and different audiences. In our hyper-connected, global ecosystem, there are so many pathways to interact with audiences. So how can you make sure your message reaches its end destination, and you don’t end up just throwing offers into the void to go unread and unnoticed? 

To boost conversion and brand relevance, our Next Best Channel machine learning model identifies the channels that customers prefer to engage on. For example, if a marketing team has been pushing an upcoming webinar mainly through email blasts that are going unread, the machine learning framework will recommend an alternative channel that would be likely to have the most engagement for each customer. 

The Next Best Channel machine learning model uses behavior data collected in Acquia CDP across web, email and social media channels to learn the transition patterns between different channels, across all customers. The model identifies individual customer channel preferences to predict both the first and second best channels to optimize future customer engagement efforts. 

Having a clear view of which marketing channels are the most active and generate the most ROI will also improve business results by helping marketing better allocate their paid ad spend and promotion efforts, so they can prioritize the platforms that earn them the highest return. Our demonstration below shows how the Next Best Channel model works within the CDP.

These models highlighted here are just a few of the hundreds of different customer parameters that can be used to build orchestrated, personalized customer engagement within Acquia CDP.. Discover how machine learning can help marketers better predict, understand and act upon the unique needs of their customers by reading our e-book: Artificial Intelligence and Customer Analytics: A Guide for Marketers.

Featured Resources

View More Resources