E-commerce companies increasingly want to offer a more personalized shopping experience for consumers. The goal is to make the shopping more enjoyable and faster for the consumer while also increasing average sales and profitability. The underlying technology that enables such personalization are so-called “recommendation engines,” algorithm-based tools that companies use to make data-driven decisions and predict what types of products or actions are likely to be of interest to a consumer.
Amazon Remembers All
When it comes to a personalized shopping experience and leveraging Big Data to make product recommendations, few ecommerce solutions can match Amazon’s technology. The sheer size of the company and the data it has at its disposal means its recommendation engine is a fine tuned machine that offers relevant and accurate product suggestions. Here’s some of the highlights of the Amazon service:
- Despite covering millions of products, the Amazon service is leveraging Big Data to narrow down to “small data” in the form of specific product. It uses data from the specific user and intelligently blends it with all users, and can “learn” over time based on which recommendations are clicked/bought/reviewed.
- Amazon is transparent about the recommendations, even asking the consumer to provide feedback. For example, shoppers can choose to “Improve Your Recommendations” by ranking purchased products which can exclude certain ones. This is especially valuable for shoppers that buy frequent gifts for others, or those who purchased a one-off item. If the shopper for example mainly buys sports equipment, but bought three books on craft beer brewing for a friend, then excluding those books as gifts will significantly alter the type (and relevancy) of recommendations.
Netflix Perfects Recommendations
Netflix is another company that has invested significant time and resources into leveraging big data to perfect their recommendation engine. The company reports up to 75 percent of content viewed through the service is based upon its recommendations, illustrating their importance in a positive user experience.
- In addition to tracking viewing habits, the service also uses what it calls“Taste Preferences”, a survey that allows the user to fine tune which genres of content they enjoy watching, and how often. This is the base of the recommendations, and also gives the user a measure of control over the recommendations.
- What sets Netflix apart is the specificity of its recommendations. It goes beyond simply offering a user “Action” or “Romance” titles and instead features micro-genres within broader categories. This makes the recommendations more relevant and gives the user the impression the recommendations are built just for them, even though the service has millions of subscribers.
Leveraging Big Data to make recommendations is so central to Netflix’s success, that the company reportedly spends $150 million a year and employs 300 people to maintain and tweak its recommendation engine. The company continues to offer a wider range of content, so relevant recommendations can ensure customers are not overwhelmed with too many choices.
Other companies utilize Big Data for recommendation engines outside of the e-commerce world, but they still provide a valuable service. Consider LinkedIn’s popular “People You May Know” feature which selects possible connections out of the company’s hundreds of millions of users. These recommendations are based on a range of data points that are designed to offer the user relevant new contacts that are based not only on shared connections but other similar interests or industry affiliations.
Some words of wisdom for e-commerce players
E-commerce companies that are looking to improve their recommendations should follow some core best practices:
- Be transparent. Amazon takes the lead here with giving users some control over their recommendations. It also explicitly spells out how it utilizes Big Data, making it a positive, and lessening the “Big Brother” aspect of data collection and its use.
- You need data. Netflix and Amazon have the benefits of millions of customers and a trove of data. Newer companies need to build up their data reserves, which can be done on the back-end and also through customer surveys and similar proactive data collection efforts.
- Make it surprising. Recommending the next Hunger Games book after the shopper bought the first book is not exactly a surprise. The best recommendation engines show the consumer something unexpected, a product that is well reviewed and likely to pique their interest.
E-commerce firms that want to increase their revenue should look at leveraging Big Data to create recommendation engines that give customers a more engaging experience. The goal is to take the customers on a journey that guides them from product to product, ideally with some sort of overall connection between all of their actions and purchases.