Published on September 01, 2021/Last edited on September 01, 2021/6 min read
The rise of predictive marketing over the last few years has sometimes raised more questions for marketers than it has answered: Do I have the right skills to look at data in a meaningful way? Do I have the right tech stack to run predictive models? What do I actually do with user predictions once I have them?
This last question is particularly important for marketers to consider as they start to add predictive insights into their messaging strategies. Predictive insights help marketers understand who to target, but they can’t move the needle without the how. Looking for inspiration? Let’s take a look at real strategies and campaigns from brands who are successfully uniting the who and the how of predictive marketing, and doing it all in Braze.
The Braze Predictive Suite, which includes both Braze Predictive Churn and the new Braze Predictive Purchases, is designed to make it as easy as possible for marketers to take action on predictive insights. By taking on the heavy lifting of complex data analysis, these machine learning models free up marketing teams to focus on what they do best: Creating effective messaging strategies that keep customers coming back for more.
It’s easy because the Predictive Suite was built with marketers in mind. Marketers don’t need to learn any new skills to set up their predictions; they just need to know which custom events in Braze relate to purchase behavior and churn behavior. Since the model is already built into Braze, the marketing team can act with total independence from other teams and technologies—no extra integrations or data pulls needed. And finally, with both the data analysis and orchestration tools all in Braze, marketers can get straight to work acting on their predictions. This proactive approach is essential for user retention and engagement.
Now, let’s explore how real marketers are taking action on predictive insights in Braze:
Braze Predictive Churn identifies users who are at risk of churning so that marketers know exactly who to target in order to increase their retention rates. Every user in the prediction is assigned a Churn Risk Score, which marketers can use throughout Campaigns and Canvases in Braze for smarter segmentation. Marketers increase user retention by leveraging Churn Risk scores to:
Understanding customer preferences has always been a top priority for Delivery Hero, a leading local delivery platform. Delivery Hero’s CRM team orchestrates their retention campaigns using Braze, but related data analysis was traditionally conducted by a separate Analytics team in a separate system. The CRM team wanted more agility to create and test their strategies, so they turned to Braze Predictive Churn. Uniting the data analysis and campaign orchestration all in one place knocked down the silos that were slowing the team down.
With Predictive Churn, the CRM team was able to create a Predictive Churn model and, over a six-week testing period, use it to better understand customer behavior and take action to improve outcomes. Most importantly, Delivery Hero’s CRM team was able to accomplish their original goal of managing the data analysis and campaign orchestration all in one place, speeding up their testing and time to value.
Braze Predictive Purchases looks at past user behaviors and patterns to predict which users are likely to make a future purchase. Each user is assigned a Purchase Likelihood Score, which marketing teams can use to accurately target users with messaging to spark quick conversions and increase campaign ROI.
Marketers can boost their revenue by using Purchase Likelihood scores to:
Health and fitness brand 8fit seeks to help users on their fitness journeys by providing custom home workouts, personalized meal plans, and self-care guidance via free and paid subscriptions on their app The 8fit team wanted to predict which free users might upgrade to a paid subscription, and to do it all within Braze, without requiring any additional tech resources.
8fit joined the Early Access program for Braze Predictive Purchases to see if they could accomplish their goals with this new feature. The first step was creating a Purchase Prediction in Braze, which assigned all of 8fit’s users with a Purchase Likelihood Score. Then 8fit used Braze Canvas to send their users promotional push notifications, in-app messages, and emails to test how accurately the model predicted which users would make a purchase. To test their hypothesis, they sent the same promotional messaging to a control group of randomly selected users.
Once the messages were sent, it quickly became clear that Predictive Purchases accurately predicted which users were likely to make a purchase: Users with high Purchase Likelihood Scores were significantly more likely to convert than those with low Purchase Likelihood Scores.
Armed with this data, 8fit decided that they could increase the efficiency and ROI of their promotional campaigns by excluding users who were unlikely to make a future purchase. This approach ensures that only users who are interested in 8fit’s services are getting messages, and 8fit isn’t wasting discounts on users who are unlikely to become loyal customers. The extra layer of personalization added by Predictive Purchases is a win-win for both 8fit and its users.
Staying one step ahead of customers and proactively anticipating their needs is a powerful way to inspire customer loyalty, so brands that leverage predictive insights in their marketing strategies have a growing competitive advantage. Use the suggestions in this article to get started, and of course, always test out different strategies to see what works best.
Ready to get started with the Braze Predictive Suite? Connect with us.