In this chapter, we continue with the following topics:
- Introduction to e-commerce
- Importance of recommender systems
- Types of recommendation methods
- Introduction to e-commerce
E-commerce is the use of electronic communications and digital information processing technology in business transactions to create, transform, and redefine relationships for value creation between or among organizations, and between organizations and individuals.
E-commerce also pertains to any form of business transaction in which the parties interact electronically rather than by physical exchanges or direct physical contact.
In 1993, Joe Pine argued that companies need to shift from the old world of mass production where standardized products, homogeneous markets, long product life, and development cycles were the rule to the new world where variety and customization supplant standardized products. He also argued that building one product is not adequate anymore. This is called mass customization, which at its core allows for a tremendous increase in variety and customization without a corresponding increase in cost. Just think of the different kinds of mobile phones and features we can get these days.
Today most companies need to develop multiple products that meet the multiple needs of multiple kinds of consumers. Mass customization has made it possible, and at the same time e-commerce has allowed companies to provide customers with more options, to showcase these customized products. However, this is a massive change to the way business is done now, compared to just a couple of decades ago. With this change, arrives a new challenge—the amount of information that a customer has to process to select a desired product has increased dramatically—a challenge of information overload. One solution to this information overload problem is the use of recommender systems.
Now we know that e-commerce is essentially the trading of products or services electronically. It is important to know that it requires other technologies such as a computer network, mobile technology, electronic fund transfers, supply chain management, and automated data collection so as to complete a feature packed e-commerce store. Most popular forms of e-commerce can be identified as:
- Online shopping: Retail sales direct to customers. Examples: Amazon and Flipkart.
- Online marketplaces: Business to consumer or consumer to consumer sales. Example: eBay.
- Business to business: Direct sales from a business to another business. Examples: Salesforce and Alibaba.com.
- Pre-retail surveys to gather product feasibility. Example: SurveyMonkey.
- E-mail marketing using promos and newsletters. Example: MailChimp.
- Demographic surveys using web contacts and social media.
The following are some sites and their primary form of e-commerce engagement:
|
Shopping |
Marketplace |
B2B |
Surveys |
E-mail marketing | |
|---|---|---|---|---|---|
|
Amazon |
Yes |
Yes | |||
|
Flipkart |
Yes |
Yes | |||
|
eBay |
Yes | ||||
|
Alibaba.com |
Yes | ||||
|
Salesforce |
Yes | ||||
|
SurveyMonkey |
Yes | ||||
|
MailChimp |
Yes | ||||
|
NetFlix |
Yes |
Yes | |||
|
YouTube |
YouTube is a typical example where the use of recommender systems is very prominent; however the actual product is not obvious. It is not a shopping site, but it is kind of a marketplace for advertisers, where the user is in fact the product. Similarly, Facebook, LinkedIn, and other social networking websites have recommender systems in place.
While we are at it, we should also think about channels for communicating with the customer. These are primarily mobile apps, SMS, e-mails, phone calls, TV ads, streaming video advertisements and newspapers, exposition, and so on. Different kinds of consumers are better reached by different kinds of communication channels. For example:
"Whereas more B2C marketers are using social channels to connect with their customers," the report said, "more B2B marketers use traditional digital channels like email."
The only channel more widely used than email among B2Bers is the corporate website, which is very or somewhat effective for 87 percent.
After all, our task at hand is to help a customer narrow down his/her search for a desired product, so we need to provide recommendations using an appropriate channel. Needless to say, today we have online services for any basic human need. We can now find entertainment, travel, food, news, and other products with just a click of a button or a phone call. This brings us to the question of the value addition by using a recommender system.
Before we decide that we want to really use a recommender system in our e-commerce site, we must be convinced that it is something that will provide valuable add-ons. In this section, we will discuss the ways in which a recommender system helps us in enhancing customer experience. A recommender system enhances the customer experience in three basic ways:
- Browsers into buyers
- Cross-sell
- Loyalty
First let's define what browsers and buyers mean. Browsers are those customers who simply browse through different products on a website. Buyers are those customers who end up buying any product or service on the website. A recommender system increases the chance by which a browser, who is just simply viewing some products, actually encounters one which he/she is really interested in buying. This converts a browser into a buyer.
A recommender system can provide product suggestions; given the fact that a customer has already selected a product for purchase. The suggestions given by recommender systems can be interesting or of the nature that those products be better bought together, with a selected product. This increases the chance of a customer buying more products from this site, rather than from a competitor's site, for example.
Loyalty time is an important factor in every decision a customer makes. Because of previous purchases or interaction with the website, a recommender system can tailor the product recommendations better. Also from a customer's perspective, it is still time consuming to visit and browse many different sites in order to find the desired product. When a customer has built trust with an e-commerce site, then it is more likely that a customer will visit that site again for a purchase. To make that happen, a recommender system can enhance the user experience by providing better recommendations than other sites.
Next, we discuss a few other terminologies that help us identify the kind of recommender system we can implement in our site. For a user to be able to find a recommendation, he/she needs to have a way to reach it. This is defined using a recommendation interface. A recommendation interface is populated using some recommendation technology, which in turn is fed using some form of input from the whole system. This is essentially how it looks:
Input → recommendation technology → recommendation interface → means to reach to a user
A recommendation interface is the kind of information that we display to the user. This can be of the following types:
Recommendation technology is the mechanism that populates the recommendation interface. All recommendation technology requires some form of input. The following are the different techniques:
- Attribute-based: Filter items based on some attributes provided by the user
- Item to item correlation: Filter items based on items similar to the selected item
- People to people correlation: Filter items based on choices of users who have similar tastes to the current user
- Aggregated rating: Order items based on user ratings
Each type of recommendation technology is fed with some data from the system. For example, in case of Amazon.com Delivers (a service by Amazon), an e-mail subscription can be set up by a user. This allows a user to receive recommendations based on some attributes of a product. The recommendation technology in this case is attribute-based, and the recommendation interface is via e-mail.
Let's see some more examples of recommendation mechanisms on a few popular sites. We will cover the following sites:
- Amazon
- Flipkart
- IMDb
- eBay
- Google News
- Times of India