Maja

Recommender systems have the three major roles in ecommerce:
 * Recommender systems &Customer support for ecommerce**

//Turning visitors into buyers (by increasing product awareness//) Companies can turn visitors into buyers by making them aware of the products available in store (mostly relevant to ones interest, sometimes for pushing new products also).This means that if a person is interested and has at the very least browsed on a particular item, an algorithm in the content-based filtering computes the similarity of that product with others and proposes the product results to the user. Another way of approaching this is by comparing a user profile and linking it to other user with similar interest, consequentially with the products they have bought or rated in a "... people who bought that also bought this..." fashion. //Increase cross sale// Recommender systems can also help increase cross sales. This is a way of retaining the customers interest and offering another product that is related to the one they are in the process of purchase for before they reach the checkout point.Schafer.B.J et al. say that the power of recommendation system in ecommerce is in creating associations between products and suggesting related products. There is an example of suggesting a fire extinguisher to a person who purchases a charcoal grill (118). //Lock in customers// The last but not least is the potential for recommender systems to lock customers in (this refers to the learning process that takes place in the recommender system as a result of interaction with the user and consequentially, the more the system learns and knows about the user`s needs, the higher the switching cost on the side of the user) (4). “Even if a competitor were to build the exact same capabilities, a customer. . . would have to spend an inordinate amount of time and energy teaching the competitor what the company already knows” (Pine et al., 1995) (Reichheld and Sesser, 1990; Reichheld,1993).

**Pre-recommeder systems organizational processes** //Database marketing// Schafer.B.J et al. talked about previous practices of small neighborhood companies to target their regulars customers. However, that is impractical and impossible to pursue in highly competitive markets with big retailers dictating norms, low employee-to-consumer ratio as well as high employee turnover rates.Some bigger companies used to simply treat their entire market as one, purely homogeneous. Others on the other hand have used postal codes, income levels, demographic characteristics to create segments of more homogeneous interests and needs among their market (121). Either way is inefficient because of huge marketing spending with bad financial resource allocation and often yield low or no results. **Recommendation systems &** **Organizational processes**

“If I have 3 million customers on the Web, I should have 3 million stores on the Web.” —Jeff Bezos, CEO of Amazon.comTM (Schafer.B.J et al.115)

//One-to-one marketing// Neumann.A in his book “Recommender systems for information providers” talks about recommender systems as a new era in marketing management. Recommender systems use a bottom up approach, rather than the so far most prevalent top-down to help organizations move on to ultimate targeting of the individual by tracing immediate needs and wants via recommender systems (3). Schafer.B.J et al.talk about the role of recommender systems in one-to-one marketing.Namely, recommeder systems exploit the power of user preferences profiling to link it to delivery and payment methods preferences in order to ensure a smooth and personalized shopping experience (122). //Ad targeting// Organizations and marketers have used ways of monitoring specific events in their current or potential customers lives.For example, if a person bought a house, he/she will start getting investment consulting from banks, housing reparations and maintenance services etc. Similarly, when someone has a newborn, advertisers start bombarding them with baby items, life insurance. Recommender systems combine this approach and see the customer as an individual as well as a part of a group. The attempt is to determine past behavior people have engaged in.Consequentially, if someone has appeared on a list for purchasing an item or service, can still be taken of the list if they have ignored several attempts to be reached with marketing messages. Some search engines use keywords entered to feed into recommender systems and show banner ads to customers based and what information they are interested.If someone searches NFL, yahoo may show them SportsAuthority.com banner ad ( Schafer et al.122). //Supply and decision support// Schafer.B.J et al.present the role of recommender systems in facilitating the work of warehouses by providing supply-chain streamlining and decision-support systems. They give predictions in terms of specific numbers about collective demand expected in say, a specific city, for a specific period of time of the year, for a specific store- all based on previously collected and analyzed data. That information is used to move the right quantity of data from wholesalers (warehouses) to retailers (stores). This translates into significant savings in the marketing budget and a more successful allocation of financial resources (116).