While the basic intent of showing recommendations is to push sales, they actually serve just beyond the better sales concept. Highly personalized content is something recommendation engines are able to deliver. This essentially means that recommendation engines on a retail platform such as Amazon are able to offer the right content to the right customer at the right time through the right channel. It makes sense to provide personalized content; after all, there is no point in showing an irrelevant product to a customer. Also, with the lower attention spans of customers, businesses want to be able to maximize their selling opportunities by showing the right products and encouraging them to buy the right products. At a very high level, personalized content recommendation is achieved in AI in several ways:

• Mapping similar products that were bought together: Let's take an example of an online shopper who searched for school bags on a shopping website. Very likely, the shopper would be interested in buying additional school-related items when buying a school bag. Therefore, displaying school bags along with notebooks, pencils, pens, and pencil cases ensures a higher probability of additional sales.
• Recommendations based on customer demographics: Showing high-end phones and stylish phone accessories as recommended products to conservative middle class customers, who generally look for steal deals, may not fetch a big upswing in sales of the recommended products. Instead, such customers might find these irrelevant recommendations to be annoying, therefore impacting their loyalty.
• Recommendations based on similarities between customers: Product recommendations to a customer are based on the products purchased or liked by other, similar customers. For example, recommending a newly-arrived cosmetic product to young women living in urban locations. The recommendation in this case is not just because of the attributes of the customer but because other customers of a similar type have bought this product. As the item grows in popularity among similar individuals, the product is chosen as the one to be recommended.
• Recommendations based on product similarities: If you search for a laptop backpack of a particular brand, along with the results of the searched item, you are also shown other brand laptop backpacks as recommendations. This recommendation is purely based on the similarity between the products.

• Recommendations based on the historical purchase profile of customers: If a customer has always purchased a particular brand of jeans, they are shown recommendations of newer varieties of jeans of the particular brand they tend to purchase. These recommendations are purely based on the historical purchases of the customer.
• Hybrid recommendations: It is possible that one or more recommendation approaches can be combined to arrive at the best recommendations for a customer. For example, a recommendation list can be arrived by using customer preferences inferred from the historical data as well as from the demographics information of the customer.

Repurchase campaigns, newsletter recommendations, rebinding the sales from abandoned carts, customized discounts and offers, and smoothened browsing experience of e-commerce sites are some of the applications of recommendation systems in the online retail industry.

Due to several prevalent use cases, it might appear that recommender systems are used in only in the e-commerce industry. However, this is not true. The following are some of the use cases of recommender systems in non e-commerce domains:

• In the pharmaceutical industry, recommender systems are applied to identify drugs patients with certain characteristics that they will respond better to
• Stocks recommendation are done based on the stock picks of a successful group of people
• YouTube and online media use a recommendation engine to serve content that is similar to the content currently being watched by the user
• Tourism recommendations are based on tourist spots that the user or similar users have visited
• Identifying skills and personality traits of future employees in various roles
• In the culinary sciences, dishes that go pair together can be explored through the application of recommender systems

The list can grow to an enormous size, given that use cases for recommendation systems exist in almost every domain.

Now that we have a basic understanding of the concept of recommendation systems and the value it offers to business, we can now move to our next section, where we attempt to understand the Jester's Jokes recommendation dataset and the problems that could be solved by building a recommendation engine.