The Secret to Doing Product Recommendations Like Amazon [Starter Guide For Shopify Brands]
Chances are, you’re already running some kind of personalized product recommendations on your Shopify store. ‘You Might Also Like’, ‘Similar Items’, or ‘Bought Together’ are examples of product discovery experiences commonly found on eCommerce sites. These strategies have consistently helped online merchants increase revenue and improve customer retention.
So, why would you bother to consider yet another personalization solution, even if it is powered by advanced machine learning technology?
The answer is simple: Not all machine learning technologies are equal.
When it comes to product recommendations, Amazon is in another league. They were the first to implement the system in eCommerce, and over a span of 20+ years, have developed a recommendation engine that now generates billions in revenue per year.
In this article, you’ll learn how Amazon got started with product recommendations, what sets their technology apart from the rest of eCommerce, and how you can begin to improve your customers’ shopping experiences with each visit to your site.
The Amazon Difference
Amazon.com pioneered personalized recommendations beginning in 1998 and has invested over two decades in machine learning research to enhance product and content recommendations across channels and devices.
The development of a machine learning-based solution is a complex, expensive endeavor – typically only pursued by those with the largest R&D budgets. It requires having a specialized team, investments into scientific research, and lengthy experimentation to train machines to achieve the desired accurate results.
Therefore, it’s no surprise that many eCommerce sites attempt to do what Amazon does with product recommendations but few, if any, are as effective.
Highly personalized content is Amazon’s key differentiator
Each time you shop on Amazon.com, pay careful attention to every detail of your buying experience. You will quickly become aware of the highly dynamic nature of their site, where almost every element of every page is dynamic.
If you switch to a mobile device, you will see that this dynamic behavior is even more pronounced for the mobile version of the site.
This even now, it turns out, applies to ads. Causing market analysts to question the impact of increasingly more ads on the customer experience:
“Advertising clouds the retailer’s ability to innovate on discovery, personalization, and any form of interactive shopping. At a certain point, every decision to improve the experience competes with lost revenue from advertising it would replace.” – Juozas Kaziukenas, Founder of Marketplace Pulse
Your ability to endlessly scroll the screen is akin to using your favorite social app; its ‘infinite’ feed introduced to keep you glued to your device.
This exercise will also help you realize that Amazon is not applying only a responsive design for the mobile version of their site.
The order of recommendations and content presentation widgets is almost entirely different. Instead of mapping the desktop version of the content to a smaller screen size, like it is the case with a responsive site design, Amazon is actually creating a mobile shopper specific version of the highly personalized content.
This is just one example of Amazon’s approach to give you a sense of how deep they go when it comes to curating the individual shopping experience.
Is the Amazon method appropriate for your Shopify store?
We know the talk about Amazon.com and that the AI technology behind Amazon Personalize sounds complex and difficult to implement. You might think that because your site is not in the same league as Amazon.com that you should not bother with the product discovery methodology they use.
The nature of eCommerce is rapidly changing. Today, most of your customers are shopping using mobile while heavily influenced by social media. This mix results in shoppers with shortened attention spans and fickle brand loyalty. It’s as much about instant gratification as it is anticipating behaviors.
Furthermore, the recent numbers across all of our eCommerce sites show that fewer than 4% of all shoppers and approximately 3% of mobile shoppers are using a site’s search feature.
All of this reveals that you can no longer rely on your site’s navigation, search capabilities, or even conventional product recommendations to effectively help your shoppers discover products of specific interest.
The only alternative is to accept the new reality and merchandise like Amazon.
How do you merchandise your Shopify store like Amazon?
The main objective of eCommerce merchandising, whether you have a small or extensive product catalog, is guiding your visitors through the buying journey. So, the act of merchandising ensures visitors get a consistent, on-brand experience; no matter how they arrive on your site or navigate around it.
In physical retail stores, merchandising relies on simple tactics applied across the entire store. The number of paths visitors can take through the store is small, shopping behavior is quite well known and changes only during seasons or special events.
For online stores, merchandising teams are faced with a complex problem – the need to apply and manage many different strategies all at once. Unlike physical stores, information about most online visitors is unknown and often changing – these visitors can follow an endless number of paths, while their preferences and behavior changes frequently.
How Amazon merchandises across the entire Customer Lifecycle:
As the first stage in a customer’s lifecycle, this is where new visitors are being exposed to your brand for the first time. They may enter your website at different pages.
At this stage of the customer lifecycle, a visitor is showing signs of interest on your site. Behind the scenes, session-based algorithms are detecting those signals of interest and modifying your recommendation strategy to create more relevant suggestions.
It’s well established, the buying stage starts when a visitor adds something to their cart. From here, your merchandising goal is to increase the average order value. You would typically do this through various up-selling or cross-selling recommendation strategies. When it comes to Amazon.com, it appears they introduce up-selling and cross-selling recommendation strategies even before a visitor adds a product to their shopping cart.
This stage of the lifecycle is about returning visitors and customers. Recommendations for the ‘Discover’ stage are dynamically populated based on general site results and incoming visitor attributes. However, as soon a visitor starts engaging with the site – with the system detecting their interests in real-time – the recommendation strategies radically alter.
And it’s not only the products that change but also the types, quantity, and ordering of the recommendation templates. As the climactic phase of the Customer Lifecycle, ‘Engage’ further demonstrates the depth and breadth of recommendation strategies available to use on your own eCommerce site.
HiConversion Recommend’s merchandising strategies connect the advanced technology behind Amazon’s personalized recommendations to your daily operations. What this really means is you can now benefit from the power of machine learning without needing to understand it.
What are you already doing to personalize your Shopify store?
Now that you know a bit about how Amazon does product recommendations, it might be a good time to think about your own personalization strategy.
If you’re already using a solution, is it part of a long-term learning project to better adapt to customers or just a way to get a few quick wins? Does it focus on just one part of your site (for example, the cart page) or does it serve the full shopping experience?
Regardless of your answers, make sure you’re tracking what really matters.
How to measure the impact of personalized product recommendations on Shopify
Before we look at an example, here are the 5 most influential metrics used to measure the impact of Amazon Personalize technology and HiConversion Recommend:
- Revenue Per Visitor (RPV) – Measurement of the amount of money generated each time a customer visits your website. It is calculated by dividing the total revenue by the total number of visitors to your site, and is a method of estimating the value of each additional visitor.
- Conversion Rate (CR) – The proportion of unique visitors that converted (i.e. made a purchase on your site).
- Cart Rate (CRT) – Indicates the rate at which items are added to the cart.
- Average Order Value (AOV) – Metric representing the value of an average order within a period of time.
- Customer Lifetime Value (CLV) – Key performance indicator that determines the value of a customer. This determines how much you can spend to recruit customers on the long term, keeping in mind the return purchases of the customer.
One of our customers, a popular US beauty brand running on Shopify Plus, noticed an immediate difference in revenue performance after applying product recommendations to their site. In fact, they saw a 101% increase in RPV compared to sessions where visitors did not engage with recommended products.
After realizing the potential, they decided to ramp up the number of product recommendation strategies across their site. The result was a disproportional increase in percentage of overall revenues attributable to product recommendations, as shown in the chart below:
While the number of visitors engaging with recommended products grew in a linear way, the percentage of revenues attributed to those recommendations grew exponentially.
Why Shopify merchants use HiConversion Recommend for better product discovery
With personalized product recommendations, customers will find more of what they like (increase conversions), visitors will stay on your site longer (higher engagement), and you’ll build a solid foundation for customer loyalty (more repeat buyers). But keep in mind – they have to be done right.
The most unique part of HiConversion Recommend is the ability to make ‘session-based’ product recommendations powered by Amazon Personalize – leveraging buying signals from web visitors, detected in real time. The video below from AWS quickly explains the need for more sophisticated, curated shopping experiences if brands want to win and keep customer loyalty:
Unlike other Shopify personalization solutions – who only serve the needs of known visitors’ preferences (roughly 10% of your visitors) – HiConversion Recommend learns and adapts to the preferences of your unknown visitors (roughly 90% of visitors).
So, what have we learned?
- Amazon has already done the hard work for us. Personalization can be complicated and resource-intensive, making it inaccessible for smaller teams. Amazon Personalize packs 20 years of learning into their algorithms so more brands can improve their product discovery without having to hire a team of machine learning experts.
- The way you merchandise is probably costing you customers. If visitors can’t find what they’re looking for quickly, they’ll likely leave your site without making a purchase. There’s a small window of opportunity to recommend products of interest and convert non-buyers to buyers.
- Great product recommendation strategies are long-term and measurable. Customer loyalty isn’t built in a day – but as Amazon has proven, good product discovery makes a lasting difference.
- Consumers are demanding highly curated, sophisticated shopping experiences. Online shopping is no longer one-size-fits-all, and in order to consistently grow this competitive market, you need to cater to individual interests and preferences.
Powered by Amazon Personalize and leveraging over 20 years of ML personalization experience on Amazon.com – HiConversion Recommend helps you focus on growing your business instead of guessing what your visitors want every time they shop on your site.