How to Build Powerful Personalized Product Recommendations

The stakes for ecommerce marketers have never been higher.

Today’s consumers now have a world of information at their fingertips that they can use to help drive their purchasing decisions. The days of shoppers simply walking into the local store and asking an associate for help are long gone. So too are the days of them looking at one or two retailers’ websites before making a decision. They now do mountains of research to identify what they want, and they know there are multiple retailers to which they can go to get just that.

So at a time when the competition has never been higher, how do you win and retain customers?

It’s all about ecommerce personalization.

Personalization has become so critical to retail marketing success because the ability to create a 1:1 experience is exactly what will foster a sense of loyalty that keeps customers coming back time and again. Showing people relevant recommendations based on their browsing history, for example, contributes to a range of benefits that have been recognized by retail marketing teams including customer loyalty, improved ROI, and increased average order value.

Personalization creates a more seamless shopping experience and establishes an emotional connection by making shoppers feel like your brand “gets them” – and a huge portion of that personalized experience for retailers is personalized product recommendations. Let’s take a look at some tips to help you make the most of these personalized product recommendations.

Understanding Personalized Product Recommendations

If you’ve ever shopped or even browsed Amazon, you have likely seen a section of recommended products. These personalized product recommendations are provided based on your existing behavior and/or your customer profile. 

For example, if you’ve bought gardening shears in the past, you might see gardening gloves, weed killer or even hoses. If you’ve browsed a particular brand, you might see the latest products from that brand. Or, even if you haven’t browsed or purchased a specific item, you might be provided recommendations based on your demographics. 

For marketers, the ability to display relevant and personalized product recommendations to customers that align with your overall merchandising strategy is key. You reach customers with products they are interested in, in turn, better connecting them to your brand and improving revenue.

Diving Into the Types of Personalized Product Suggestions

There are several types of personalized product recommendations that can be served to customers. Here are a few different types of recommendations you should employ: 

• Next Best Purchase: You could recommend the next product a shopper is likely to buy based on shopper data, behavioral data and your product catalog. 
• Interaction History: Act on signals from each shopper’s interaction with your product catalog, like views, purchases, searches and cart adds with recommendations that change with each of these interactions. 
• Co-Recommendations:  By leveraging what you know about other shoppers who have similar purchase patterns, you can recommend products that your shoppers might be most likely to buy next based on those similar shoppers’ actions.
• Best Sellers: Put your best foot forward with the power of your product catalog to surface highest selling products to shoppers based on how many were purchased in the last day, week or month. 
• New Arrivals: Some shoppers love seeing (and buying) the newest items you have to offer. Show your new arrivals shoppers all the latest and greatest from your product catalog.  

Choosing the right type of product recommendation will be dependent on your available customer data, your products and your goals. 

Let’s take a look at some tips to help you make the most of these personalized product recommendations.

Effective Delivery of Personalized Product Recommendations

The benefits of creating a personalized, 1:1 retail marketing experience are clear, but how do you actually get there? If you’re not exactly sure, don’t fret – you’re not alone. A study by Forrester found that only 12% of marketers feel they are “very effective” at delivering personalized experiences to customers. 

In the same breath, this means that your organization can break away from static or segmented product recommendations. Moving to implement personalized product recommendations with the help of machine learning, you could stand to gain a competitive advantage.

To start, you need technology to help you track and tie together customer data, behavioral data and product data. From there, you can begin to understand your customers’ past actions and their predicted future actions.

With those capabilities in place, there are numerous ways to approach personalization. For example, you might run targeted campaigns for different audience segments or include personalized product recommendations within batch or triggered emails.

Today, let’s focus on best practices for the latter of those two options — how to develop a strategy for effective personalized product recommendations. 

1) Identify Your Audience

First and foremost, you always need to ask yourself “who are the customers that will ultimately receive this campaign?” The answer to that question should be a driving factor in deciding how you approach personalized product recommendations.

For example, a customer who came to your online store and abandoned it at a category-level page without ever clicking on a specific product is much less engaged than a customer who added a product to their cart and then abandoned it.

With that in mind, you should use a broader product recommendation strategy for the first customer (the one who abandoned a category-level page), such as showing best sellers within the category they abandoned. You can get more targeted for the second customer (the one who abandoned their cart), for instance by doing a co-view recommendation that features products most viewed by other customers who also viewed the carted product. 

For instance, observe how Hammacher Schlemmer recommends products to customers within emails.

This customer in particular began searching for outdoor seating. As a result, they receive suggestions for beach and yard chairs as well as some additional yard products.

When customers explore the website, the retail personalization software tracks their behavior to determine which products they’re interested in as well as which products they’ve never seen but may have an affinity for. This information enables each customer to receive personalized product recommendations based on their interests within the same marketing campaign.

We can see this in the second example from Hammacher Schlemmer, which comes from the same email campaign as the first, but features an entirely different set of products — this time car accessories. Critically, both examples show a variety of products within a specific category to encourage continued engagement by exploring other related products.

2) Test Campaigns Regularly

Once you put in place an initial strategy around personalized product recommendations, let it run for a minimum of 2-3 weeks so that you can collect some baseline benchmarks for your specific customers and campaigns. Taking the time to collect these benchmarks is very important, as each retailer’s customers engage differently based on factors like vertical, products, price point, etc.

Once you have those baseline benchmarks, you can begin to test different strategies for personalized product recommendations to see if you can improve on your customer engagement metrics. This testing is an extremely important part of regularly improving your email marketing and is something that you should repeat regularly.

Far too often, marketers test one theory, choose the winner and then move on. The problem with this approach is that it doesn’t go far enough. You should always be testing at least one variable in all of your recurring campaigns, and your product recommendation strategy is an easy one to test. It’s important to keep in mind that there is no “magic bullet” that works for all retailers, so constantly testing and optimizing should always be a part of your efforts to get personalized product recommendations right.

3) Trust Your Data

One trap that even the strongest marketers fall into is being too close to their own business. When you live and breathe your brand, it becomes second nature to assume that you understand everything about your customers. But this assumption isn’t always right and it often leads marketers to over-engineer the rules for personalized product recommendations.

For instance, thinking “we know this person bought shirts, so let’s recommend shorts” makes perfect sense, but it’s a decision based on very limited input. The beauty of using a platform that collects large quantities of data and automates product recommendations in real-time based on customers’ onsite behavior and predicted affinities is that the technology can make brand or category associations we would never even think to put together. Sometimes it’s simply best to let the data do all the heavy lifting.

Honing Your Personalized Product Recommendations is an Ongoing Effort

The bottom line here is that honing your strategy for personalized product recommendations is not an exact science. What works today might not necessarily work six months from now, so it’s something you need to iterate regularly. And given that providing relevant content to your customers is a never-ending process, it’s best to find the most efficient ways to meet that need.

Ready for more? Check out how Hammacher Schlemmer achieved 16% year-over-year revenue growth in email by activating customer data to power intelligent personalized product recommendations.