5 Key Lesson Learned From Amazon’s Product Recommendation System

June 17, 2024

November 16, 2022

•

6 min read

Amazon is without a doubt the shining star of the eCommerce world. Its ability to guide shoppers to the products they want is so effective that 20% use it at the beginning of their shopping journey. That’s second only to Google!

One of the reasons its eCommerce wing is well-trusted by consumers is down to its powerful product recommendation system. The company pioneered the notion of recommendations that are super personal.

We can always look at the big players to teach us a thing or two about success. So, in this article, we’ll pick out the top lessons you can take away from Amazon to boost discovery through product recommendations.

The Foundation Behind Amazon’s Recommendation System

(Source: Amazon)

Before we jump into the key takeaways from Amazon’s recommendation system, let’s first get some context.

Amazon’s recommendation system is centered around personalization. Since the 2000s, the company has invested itself in creating a powerful recommendation engine. One that mimics the pre-eCommerce era of word-of-mouth recommendations from family and friends.

As Brent Smith of Amazon wrote alongside Microsoft’s Greg Linden

“...discovery should be like talking with a friend who knows you, knows what you like, works with you at every step, and anticipates your needs”.

This approach to discovery is the backbone of the company’s recommendation engine. Every step of a consumer’s shopping journey has personalized product recommendations woven into it. These recommendations work so well that 30% of its revenue comes from them. By taking this stance, the company has expertly humanized the online shopping experience.

So what goes on behind our screens to make this approach to recommendations feel so personal? It’s achieved with an algorithm known as item-to-item collaborative filtering.

(Source: Medium)

This type of filtering diverges from the traditions by matching each of a shopper’s purchased and rated items to similar ones, rather than simply matching similar customers. The algorithm then decides which products are the most relevant by finding items customers usually purchase together. It’s this item-first approach that gives Amazon its edge.

Furthermore, the algorithm is able to scale independently of the number of customers and items in a product catalog. Hence, Amazon can present high-quality product recommendations in real time regardless of its giant inventory.

Humanizing the online experience is something that all eCommerce merchants should strive for. Product recommendations with a personal touch similar to Amazon’s are an excellent starting point.

Fortunately, you don’t need to have the big bucks and decades of experience to achieve this. Here are 5 lessons you can learn from Amazon and implement right away in your Shopify store.

5 Key Takeaways From Amazon’s Product Recommendation System

#1 - Leverage data to hyper-personalize product recommendations

A crucial step in delivering tailored product recommendations is data. This is the bread and butter of Amazon’s recommendation engine. You don’t have to browse through Amazon for very long to see that the website is taking notes.

It’s not just your browsing history that gets examined. A visitor’s location, recent purchases, saved items, and trending products play a part too. It’s the use of this robust data that makes Amazon’s product recommendations feel like they’ve read your mind.

To deliver product recommendations that are totally in tune with your customers, there are a couple of things you need to do.

If you’re a Shopify merchant, seek out an AI-powered product recommendation app. These apps use AI to track user behavior and surface relevant recommendations. They can also keep tabs on trending products, and decipher what items are frequently bought together. Wiser, Limespot and Frequently Bought Together are great options for Amazon-like recommendations.

To be able to get really personal, the other vital task is to get your customers to sign up. This is the key to recommendations that are on a more granular level. You’re able to tap into the items they’re saving or favoriting, and see the full scope of the products they’re buying and viewing, in addition to personal information. This enables you to build a richer product recommendation experience.

Incentivize shoppers to sign up with the promise of a special offer or discount. Alternatively, you can take a page out of Amazon’s book and plainly tell them it’s the gateway to better recommendations.

#2 - Harness the power of social proof

Whilst browsing on Amazon there’s something you’ll almost always see beneath the product thumbnail - star ratings.

The modern shopper is becoming a little numb to claims made by brands. Whilst product descriptions are a decent source of information, your customers need more. The authenticity of social proof such as star ratings adds credence to your recommendations.

Seeing dozens or even hundreds of positive ratings beside the products “You May Also Like” makes this phrase so much more believable.

If product reviews aren’t something you’re collecting yet, get on that train now! There are tons of great review apps like Yotpo, Judge.me, and Ali Reviews that will assist you in collecting and showcasing reviews when you recommend products.

Another interesting form of social proof that Amazon uses is showing what people buy after viewing or searching for a particular item.

Not only is this great for the uncertain shopper, but it’s also hugely beneficial for you. This strategy places a focus on products that are most likely to result in a conversion. As a result, you’ll push shoppers further along the purchase funnel.

#3 - Use product comparisons

Another golden online product recommendation tactic utilized by Amazon is product comparison.

When you click on certain product pages, after a few scrolls you’ll see a row of products ready to be compared side-by-side. What's more, the perfect product can be added to the cart immediately.

This is excellent as, at a glance, shoppers can see how similar items stack up against each other. With vital information such as price and technical aspects right before their eyes, it becomes much simpler to decide what’s best.

The whole point of product recommendations is to help customers find the right item easily and quickly. Allowing for product comparisons hits this nail squarely on the head.

This approach to recommendations works especially well for stores that sell complex products like electronics and home furnishings. That being said, stores with more day-to-day offerings can benefit from this too.

If you sell advanced or technical products, pick out some key attributes that will assist shoppers in decision-making. For other products, simply including the price and star rating of recommended products works just as well for a comparison.

#4 - Recommend products at every stage

As Amazon’s Brent Smith wrote, the facilitator of product recommendations should “work[s] with you at every step”, and they sure do on Amazon.

There is literally not a page you can click on without seeing some sort of product recommendation. The homepage, category pages, product pages, add-to-cart pop-up, cart page, and post-purchase email all include recommendations.

Recommendations on an Amazon category page.

What this does is really expedite the product discovery process. It ensures there is no point where shoppers feel lost or uninspired as they traverse Amazon’s huge inventory. It also catapults the chances of scoring a higher AOV.

What’s impressive, though, is that the company does this without ever feeling overbearing. The recommendations are so well-tailored to a customer’s situation that it instead feels helpful.

Shopify merchants can adopt this technique with Wiser and Limespot. Both apps allow you to display recommendations on all of your Shopify pages as well as in emails.

When adding recommendations to the pages in your store and in post-purchase emails, it’s important to consider the state of mind of your customers. Let’s quickly break that down:

Homepage: At this stage, you’re not quite sure what a visitor is looking for, and maybe they aren’t either. Hedge your bets on trending items, new arrivals, and the hottest deals.
Collection page: They’ve got a slight idea of what they want so top-selling items in the category work a treat here. If they’re a returning customer, similar products to what they’ve purchased within the category are a good idea, as well as recently viewed.
Product page: Items that are very similar to what is being viewed should go here. The type of product they want is becoming clearer to them. Additionally, at this point, your customer may not be completely sold on what they are viewing yet. You can sweeten the prospect of the product being viewed by recommending items that would turn it into a set or an entire outfit.
Cart page: Your customer knows that this is what they want, so now it’s time to cross-sell. Products that are frequently bought together are right at home here, as well as items that add value to the item/s in the cart.
Post-purchase email: Something we’ve seen Amazon do at this stage is to encourage shoppers to explore more from the brand they purchased.

It’s clear at this point that your customer believed in the quality of the brand. So, if you stock items from various vendors, consider doing this too. Alternatively, you can recommend new arrivals in the category they purchased from, or products that act as an add-on or accessory to what they bought.

#5 - Vary the types of product recommendations

Following on nicely from the previous takeaway, another very noticeable feature is the variety of recommendations seen on Amazon.

Amazon goes beyond frequently bought together and related items. You’ll see product recommendations based around deals, vendors, personalized ones based on purchases and browsing, featured items, newer versions, best-selling items in the category, and even what’s trending in your area. This is all thanks to the effective use of customer data.

As we mentioned before, Amazon makes recommendations everywhere, and a key reason for that not feeling oppressive is variety.

It’s a brilliant tactic that eCommerce merchants of all sizes can experiment with. Think about how you can leverage your customer data to produce creative product recommendations. Consider the different customer personas you have. What type of recommendations can you build exclusively for them?

You can also think about any expertise you or your employees may hold that can give your store a truly unique style of recommendations. For instance, if you work in fashion, maybe you could include hot picks as chosen by your Fashion Buyer.

Build Your Product Recommendations Like A Pro

Delivering relevant and innovative product recommendations is something all eCommerce merchants need to aim for. It’s the key to creating an effortless and enjoyable shopping experience.

Not every shopper knows what they want when they land on your Shopify store. So, it’s your job to be their masterful guide. Taking notes from a big player like Amazon will have you on your way to being an expert recommender.

Doki Nguyen

•

Content Marketing Specialist

November 16, 2022

•

6 min read

There's More

5 Key Lesson Learned From Amazon’s Product Recommendation System

Shopping shouldn’t feel like a chore

Related Posts

Level Up Your eCommerce Game With Targeted Email Marketing

Expert Opinion: AI & CRO | TAG Webinar Roundup

Expert Opinion: AI & CRO | Brancoy Webinar Roundup