Every time one of Amazon’s 150 million Prime members visit Amazon.com, they see features on the homepage like “Inspired by your shopping trends” and “Based on items you viewed” with carousels of additional products to browse and, ideally, to buy. These aren’t random items picked out by the platform. Instead, these examples show a powerful tool for e-commerce marketers being deployed in an incredibly strategic and intelligent way. AI-powered product recommendations can radically reshape the way consumers shop, in a way that traditional manual recommendations can’t come close to touching.
What Are AI-Based Product Recommendations?
A recent study from the research firm Orbis Research described AI-based recommendation engines as “data-filtering tools that make use of various algorithms and data to recommend the most relevant items to a particular customer.”
To do so, they pull in data from prior customer behavior (both from the current customer and from their entire customer pool), like searches, clicks, and purchases. Then, they calculate what will appeal most to that particular consumer in the future. AI-driven product recommendations help ensure customers find products they want to buy quickly and easily. They also allow brands to highlight the products other customers love most and get those products in front of new consumers. It even creates opportunities to cross-sell and upsell.
In Amazon’s case, AI-driven product recommendations fuel discovery, which Martech Advisor editor Shabana Arora pointed out, is particularly important for so-called “long-tail items,” or those that aren’t particularly popular.
“Recommending long-tail items to shoppers is critical because, if successful, it has the potential of giving ROI on slow-moving inventory,” she wrote.
Streaming service Netflix is another great example of a brand tapping into AI-based recommendations. According to Arora, the platform has spent years tweaking its recommendation algorithms in order to highlight the best possible content for each of its 180 million subscribers. As a result, Netflix users don’t have to scroll endlessly to find something to watch, which, Arora pointed out, translates to increased viewership and decreased member churn. Netflix has even said its AI-based recommendations save the company $1 billion a year.
For more on the opportunities of AI:
Why Are AI-Based Product Recommendations So Important?
Product (or, in Netflix’s case, content) discovery is one of the main reasons brands should offer AI-based product recommendations. According to management consulting firm McKinsey, 35% of Amazon purchases — and 75% of Netflix content — hailed from these product recommendations back in 2013.
But, as Arora pointed out, they also enhance the user experience by offering the right products to the right customer at the right time. (And isn’t that the age-old mantra of digital marketing?) According to her figures, roughly 35% of all sales were generated by recommendation engines as of 2016. That number is undoubtedly higher now.
In recent years, the retail industry, in particular, has embraced AI-based product recommendation engines. The Orbis Research report found the increase in information on the Internet, combined with a rise in the number of users, has made it vital for retail brands to search, map, and provide consumers with “the relevant chunk of information according to their preferences and tastes.” In other words, don’t make your customers do the heavy lifting — instead, offer up relevant products on a digital silver platter.
As a result, Orbis projects the recommendation engine market, which was valued at $1.2 billion in 2019, will grow over 34% from 2020 to 2025 based on an increased need for customer retention and return on investment. It’s also being driven by growth in e-commerce more broadly, which, in turn, is increasing demand for recommendation engines that enhance the customer experience and fuel retention. Data from market research firm IDC is in alignment. It found retailers invested nearly $2.5 billion in product recommendation systems, automated customer service agents, and shopping advisors last year alone.
It used to be a much heavier lift. As this CIO article points out, brands used to need internal expertise in machine learning in order to build personalized recommendation systems with deep-learning models. Now, however, plug-and-play AI recommendation engines not only help brands adapt offerings to consumer behavior, they also help retailers adjust for factors like what’s in stock in real time.
But, of course, to make the best possible recommendations, retailers need as much customer data as possible. That’s why some are also tapping into insights from product reviews and interaction with customer service agents. Another CIO article on product recommendations says next-generation product recommendations have to go a step beyond just the products a customer might like and incorporate all the data a brand has about a given consumer. What’s more, they’ll have to continue to update those recommendations based on customer behavior at that very moment. That means finding the right platform is critical.
But brands also have to be careful not to make the wrong recommendations, which could yield a bad customer experience. Consider, for example, if a brand recommended a product that was out of stock. Or think about how you’ve felt in your own experiences with brands when they’ve repeatedly suggested a product you don’t want. That, CIO says, is preventable, but you have to have an accurate picture of each consumer across all touchpoints.
For current retail industry predictions:
AdRoll’s AI-Driven Product Recommendations
AdRoll is among the platforms offering a solution for AI-driven product recommendations. But it’s not just any solution.
AdRoll uses data from every site visitor to drive product discovery. We save 30 days of consumers’ viewed products, which we feed into our recommendation engine. That helps brands capture customers on the fence about a purchase by making it easier for them to find products of interest over and over again.
We also help clients set up their product feeds. This enables brands to customize how their recommendations appear so they match brand designs, and fit seamlessly into brand layouts. AdRoll’s AI-driven product recommendations use dynamic ad, email, and online store content to create personalized cross-channel experiences proven to convert. Clients can customize product recommendations to display previously viewed or recommended products – and, as a result, stay top-of-mind throughout the customer journey, no matter what shape it takes.
For more growth marketing strategies for challenger brands:
AI-driven product recommendations are crucial to winning consumers’ hearts and minds because they offer personalized recommendations to customers and prospects whose expectations are only going higher. These product recommendations make shopping less impersonal and help brands gently remind shoppers what they have seen and loved to encourage future engagement. By offering up those products, brands increase the likelihood of conversions — and even average order value. Try AdRoll’s solution out for free here.
Laura Smous is the Senior Director of Product Marketing at AdRoll.