Good cross-sale recommendations can help you increase your customer lifetime value and build loyalty with your customers.
How many promotions, ads, and offers do you see or hear every day?
If you’re like most of us, the answer to that question is more than five thousand.
With so many commercial messages thrown at us so often, we’ve all developed our own filters for ignoring, deleting, and archiving anything that we don’t immediately find relevant to us.
As more and more brands fight for their customers’ attention, we’ll all be stricter and more discerning in what we ignore, which means that businesses that fail to deliver personalized and valuable messages to their customers will quickly find themselves ignored and forgotten.
One of the best ways to make sure your customers pay attention to you is to target them with only the most valuable and relevant messages that are applicable to each one of them, individually. If you’re an ecommerce operator, a great way to stay relevant is to identify the one product that each of your customers wants to buy next and target them with a personalized message to get them to purchase that product.
For most retailers, repeat customers are the lifeblood of their business. According to an Adobe Digital Index Study reported by Payhelm:
It’s no surprise that a customer that keeps coming back is a customer you should hold on to. So the main question you’re probably asking yourself is “how do I keep a customer coming back?”
If a customer has a good experience with your brand, their checkout and shipping experiences are seamless, and they receive good customer service, then getting them to come back can be as simple as reminding them that you exist and offering them a new product that they may be interested in. That’s where cross-sell recommendations come in.
Product recommendations can come in many forms. You’ve likely seen in-store recommendations that show up on a product page when you’re shopping online. Sometimes these recommendations show up as a list of products that other customers have purchased after browsing the same product you’re browsing. Other times they show up as a list of products on the homepage or account page of your account.
At Apteo, we focus on cross-sale recommendations in the context of customers who have already purchased from you. When a customer buys a product from you, there’s a good chance they’ll be interested in other products that you sell. By getting customers to come back and make a follow up purchase, you’ll start to grow your customer lifetime value, increase retention, and increase repeat customers, helping you reap all of the benefits mentioned above.
If you can properly identify which of your products an existing customer of yours would be most interested in, you can send your customers a targeted, personalized message that has a higher-than-average probability of getting them to make a purchase. And if you’re using the right tool, you can also optimize the content, frequency, and even discount amount for each customer, helping you grow sales while maximizing your profit.
So while the benefits of good cross-sell recommendations are clear, it’s not immediately obvious how to create and deliver these recommendations.
When focusing on creating recommendations for customers who have already made a purchase, it’s important to only send out campaigns that have a high probability of converting. If you send out campaigns that customers don’t care about, you run the risk of customers ignoring your marketing campaigns, or worse, sending your emails to spam, which could hurt your sender reputation.
So how do we know whether a customer is likely to buy another product, and if so, which product they’re most likely to buy next?
This is where all of the historical data that you’ve collected on your customers can really shine. By properly mining that data, you can find common patterns of behavior among your customers and use those patterns to incentivize customers to make another purchase. While it might be possible to find some of these patterns in Excel, today, it’s best to use A.I. or machine learning software to do it for you.
While that may sound intimidating, today you don’t need to be a data scientist to use A.I. There are a variety of tools out there that analyze and organize your data, making it easy for you to get back to growing your business. But even if you use software to find these patterns for you, it’s still helpful to know what kinds of data and attributes go into finding these patterns. Knowing the inputs can help you better contextualize the outputs, and can help provide you with new and better ideas for growing your business.
When it comes to predicting what someone is likely to do next, there’s no substitute for knowing what they’ve done in the past. Behavioral data is ideal for machine learning algorithms that are optimized to predict what a customer will buy next.
What we’ve found is that having a customer’s purchase history can provide all sorts of valuable little tidbits of information that can be helpful when training a model. This can include any of the following:
While behavioral data is the golden standard for creating optimal machine learning models to predict cross-sells, demographic data can also be extremely useful. A customer’s location, gender, income, even email domain can all provide slight clues that, when combined with each other, can help an algorithm figure out important purchase patterns.
Third party data, such as which sites or apps a user visits, what they’ve spent on other products, or what products they’ve purchased on other websites can all be helpful. Some of it can be categorized as behavioral data, in which case it could potentially be extremely useful in predicting a cross-sale. Other times, though, this data is anonymized and aggregated, making it less useful when it comes to understanding what a customer will do next.
While it may be helpful to include this data in a model, it’s best to use it as supplementary information and to not rely on it as the primary method for understanding what a customer will do next. Similarly, aggregated data across multiple users or customers can be useful, but because it’s aggregated across multiple customers, it loses a bit of predictive power.
It’s always best to use individualized behavioral data whenever you can. When it comes to behavioral data, zero-party and first-party data that you’ve acquired is invaluable. At Apteo, we use a wide variety of first-party data to analyze your customers and predict what they’ll buy next.
Now let's assume that you’re not using a service to create your cross-sell campaigns, but instead, you’re structuring all the data yourself and feeding it into a machine learning algorithm.
How would you actually do this?
At their core, A.I. and machine learning algorithms are basically just giant pattern-finding tools. You give them a bunch of examples about what has happened in the past and associate those examples with what happened in the future and they’ll break everything down for you and remember the patterns they find.
At Apteo, we create lots of examples of what your customers have done in the past and what products they buy next (or, we see that they didn’t buy another product and we use that as an example that our A.I. can learn from), and we feed that all into a giant algorithm based on gradient boosted trees and use the results to predict what each of your customers do next.
As your customers go through and interact with your site, making new purchases, or as time goes by and we see that they haven’t come back, we’ll update our predictions continuously and segment everyone into groups of customers that are likely to buy a specific product next.
You can then use these segments to send out targeted email messages, incentivizing your customers to come back and make a purchase. If you use Apteo’s dynamic cross-sell campaigns, we’ll also go through and optimize your email frequency and discount amounts for you.
A good cross-sale campaign can significantly grow a retailer’s sales - Amazon claims that cross-sales make up 35% of their revenue. While you may not have the scale and impact that Amazon does, a good cross-sale campaign is free to execute on and can reap huge rewards in terms of customer loyalty and lifetime value.
As retailers continue to struggle with growing acquisition costs and the impact of Apple’s iOS privacy changes, cross-sale campaigns should play a key role in a highly personalized, targeted marketing strategy. The more your customers feel like you’re speaking their language, the more they’ll be likely to interact with you, and the more they interact with you, the more likely they are to buy your products.
If you’d like to get cross-sale recommendations, retention analytics, product journeys, bundle analytics, and automatic segments completely for free, you can sign up for a free account at Apteo and get started with personalized marketing in just minutes.
Shanif Dhanani is the co-founder & CEO of Apteo. Prior to Apteo, Shanif was a data scientist and software engineer at Twitter, and prior to that he was the lead engineer and head of analytics at TapCommerce, a NYC-based ad tech startup acquired by Twitter. He has a passion for all things data and analytics, loves adventure traveling, and generally loves living in New York City.