Where to Use Predictive Scoring Throughout Your Customer Journeys

Where to Use Predictive Scoring Throughout Your Customer Journeys

If you’re a marketer in 2020, there’s no doubt you’ve heard the benefits of AI and machine learning. And while AI is a game-changer for marketers looking to do more with their budgets, data, and technology, we often focus on popular topics, like recommendations, leaving one of the most brilliant uses of AI for marketers often goes untouched: Predictive Scoring. This might even be your first time hearing that term. Basically, Predictive Scoring is: 

Predictive scoring, much like traditional lead scoring techniques, uses AI algorithms to analyze every behavior a customer (both known and unknown) has exhibited with your brand and assigns a value from 1-100 that shows, comparative to others, how likely any given customer is to engage, convert, or churn.

Pretty amazing right? Predictive Scoring can help you stop wasting your time and money sending the wrong customers the wrong messages, and really hone in on what will make the most impact for your organization. Curious to see how Predictive Scoring is used in practice? Here are some of the most popular campaigns Blueshift customers have created with the help of Predictive Scoring.

Likelihood to Churn: Subscription Renewal Campaign

Let’s take a look at how an e-learning brand could easily use Predictive Scoring to revamp their existing Renewal Campaigns.

Many e-learning companies operate off of a subscription model — this might take the form of a yearly subscription to unlimited classes or just a subscription for a specific class or term length. Regardless, there comes a point when the customers decide to renew or move on. Marketers need to get messaging up to the subscription renewal just right to keep the majority of their students as paying customers. A great way to tailor messaging, and ultimately save your brand money, is through “likelihood to churn” scores. These Predictive Scores calculate how likely a certain customer is to not renew, or churn, on a scale of 1-100.

Campaign Tip: We can use these scores to build segments or personalized which messaging and offers customers receive. Customers with a relatively low likelihood of churning (1-40) can receive standard messaging, simply reminding them that it’s time to renew their subscription and what they’ll miss out on by not renewing. For folks in the midrange of churn likelihood (41-79), you can add more targeted messaging and possibly a small incentive such as 5-10% off their subscription to get them converting. And lastly for the folks who most likely won’t stick around (80-100), throw out your highest value offers (10%+ off) to get this segment of folks converting and work further on turning them into brand loyalists with solid recommendations and content throughout their next subscription period.

Likelihood to Purchase: Abandoned Cart Campaign

Let’s take a look at how a hospitality brand could easily use Predictive Scoring to revamp their existing Abandoned Cart Campaigns.

We’re all been there, gabbing with your friends and family leads to dreaming of fun future vacations — and one of the first things we do is check flight and hotel prices. What separates daydreaming from confirmed plans can often be achieved through solid abandoned cart campaigns.

An “abandoned cart” describes when a customer adds something like a flight or hotel room into the hospitality site’s check-out system without completing a purchase. And while these campaigns are primarily associated with ecommerce, there is a huge benefit to different verticals using these campaigns (and adding Predictive Scoring into the mix). To help brands optimize their abandoned cart messaging and offers, “likelihood to purchase” scores are a fantastic start.

Predictive scores graph

Campaign Tip: Rather than spamming potential guests or flyers with messages, hoping one sticks, Predictive Scoring can help marketers take a much more targeted approach to get customers back to their carts. For those customers who have a higher likelihood of purchasing (51-100) send them a standard “come back” coupon or offer combined with persuasive messaging and recommendations. But for the customers who are less likely to purchase, (0-50) send a more exclusive offer, combined with Predictive Channel Engagement Scores to really ensure they’re seeing your message (and are aware of the great deal they have sitting in their inbox, push center, etc.)

Likelihood to Engage: Welcome Series

et’s take a look at how a streaming service could easily use Predictive Scoring to revamp their existing Welcome Series.

Welcome series put a ton of pressure on marketers — and for good reason, it’s the customer’s first glimpse of your brand, and humans are known for sticking to first impressions. Using Predictive Scores like “likelihood to engage” can help marketers determine how often message new customers within the campaign.

Campaign Tip: For customers who are already familiar with and excited about your streaming service — and with a higher engagement score (51-100), they’ll want more communications with recommendations, direct links to stream, and the most ways to engage with your brand as possible. For those who show a lower propensity to engage (0-50), more educational content can be a great alternative.

Likelihood to Add to Cart: Recently Viewed Products Campaign

Let’s take a look at how an e-commerce brand could easily use Predictive Scoring to revamp its existing Recently Viewed Product Campaign.

More than ever before, e-commerce is relying on 1:1 recommendations to keep shoppers browsing and engaging with messages. The gold standard comes from brands like Amazon who have achieved a seemingly endless scroll of tailored content (both on their site and within marketing messages). Predictive Scores can help your brand be more “Amazonian” and help get customers who are soon to convert over the edge with “likelihood to add to cart”.

Campaign Tip: For customers with a high likelihood of adding an item to their cart (51-100), they’re probably reaching the end of their buying decision and most likely know what exactly they’d like to purchase. These customers should receive messages recommending products they’ve viewed the most and most recently, with maybe one or two very similar options to consider. For the folks who are just starting out their decision process and are less likely to add to cart (0-51), broad recommendations based on their past views and offering up a wider variety of products is how you can best help them continue their journey to conversion. All of these targeted messages can also be carried over to Syndications, or advertising across the web, with each segment getting targeted with varying levels of frequency.

Want to learn more about how Blueshift’s Predictive Scoring (or maybe just AI Marketing in general) works? Check out our Ultimate Guide to AI Marketing to learn more, or see Predictive Scores in action with one of our AI experts today.