Dany_jessie


 * Introduction**

All of the managers’ activities within an organization involve decision-making. The task of reaching the best possible decision is very difficult and requires a lot of information gathering. Thus, both information systems (IS) and technologies (IT) are integrated to corporate culture because they help managers make the smarter choices faster.

It is important for managers to understand how IT differs from IS and acknowledge that a company can be successful even though it is not the most technologically advanced. IS occurs through the incorporation of data and communication to generate knowledge. Swapping post-its between two persons can be considered an IS. In contrast, IT is simply the mechanical, technological products that serve the purpose of IS. In this case, the pen and paper alone would be the IT; showing how IT alone does not have the capacity to select the optimal choice. For example, software (Windows Vista), hardware (laptop), networks, etc. are all IT. Therefore, organizations use IT as a method to collect, process, accumulate, and distribute information within the business; as well as suppliers and customers. IS concerns the fusion between human and machine interactions that lead to expert decision-making.

The type of technology a manager decides to implement can either help or hinder a company’s operations. There is a plethora of existing information systems that serve a variety of purposes. This report aims to explain the functionality, data flow, process flow, and implementation of recommendation systems.

Recommendation systems are a high potential IS. Its implementation may increase customer retention and satisfaction, sales, and profits culmination in business success. These systems have increasingly been used by organizations to acquire data on customers, foresee market trends, and increase inventory fluidity. Especially in electronic commerce (e-commerce), these systems are disseminated amongst online shopping websites. This IS can generate consumer profiles based on past purchases, research related products, and common habits to in turn, display information of interest to users. Whether it be in digital television, online query services, library catalogs, or Internet shopping carts, recommendation systems are capable of presenting managers and customers with relevant information that become knowledgeable decision-making.


 * Implementation**

//i.Strategies//

A strategy consists of plans with the best course of actions that achieve the organization's objectives and goals. These plans include both long and short-term procedures. All activities in a company should be supported by strategies, including the implementation of Recommender systems.

A first strategic approach to Recommender systems is a comprehensive and interactive graphic user interface. The technology should not require extensive training for users to input and retrieve data into the system: the design must be simple and efficient. When the user performs a query, the results have to pertinent and timely. Thus, the plan of action is to ensure that the interface is self-explanatory and the system's software is accurate and agile. In e-commerce, GUI suggest to shoppers other products and services that correspond to their desires and requirements (Felfernig et al, 2008).

For example, on Bestbuy.ca, once the customer is viewing a product in detail, the sidebar displays other products that either relate to the one being seen (like complementary accessories), to the search words (eg: “laptop”), or to other goods that different customers purchased together with that product. The sidebar recommendations include name, image, link to the product details, and an option to add it to the cart. This GUI is an effective strategy because it captures the customer's attention with a visual that gives a brief scope on what it is and already allows shoppers to add it to the cart without learning the price.

Another important strategy is to guarantee the software is always up-to-date. Technology is in constant evolution and it is important that the system is upgraded to sustain level of quality of services. For example, as the Internet providers become faster, the recommendation system should be able to display results of similar goods or services related to the research in the same time frame. This strategy enables the system to remain relevant in the future.

Another strategy that may facilitate the integration of a recommender system with users is to release a beta version of the actual system first. This test version of the technology allows enough time for end users to familiarize with the system before it is completely implemented. Also, users should be able to comment or even be officially surveyed on the system's usability. The opinions point towards system mistakes and qualities, as well as give extended opportunity for the organization to adjust and improve the IS.

For example, Orkut, Google's social networking site, has a feature that suggests other profiles to users depending on their own social circle. The website recommends at first those who have the highest number of friends in common with users. Nonetheless, since 2009, the feature was in development, as well as the whole website is constantly updated. Users may choose to use the beta version or switch back to the old until they have time to readjust to new grounds. At first, Orkut itself suggested people, but in the newer version, other profiles can suggest individuals to the users.

//ii. Tactics//

The main tactic for successful implementation of a recommender system as a marketing strategy is a graphic user interface. It is pointless to adopt this type of system if the end user won’t understand or access the recommendations provided. A study conducted with users of MovieLense recommender system proved that users are more satisfied with the recommendations they receive when they are familiar with how the system works (McNee et al. 2003). The GUI ensures that the recommended items are easily visible to the potential consumer through a well designed website. It should be designed respecting Jakob Nielsen’s principles of usability, requiring minimal effort to operate (Nielsen & Levi, 2004). This ensures not only that the buyer sees the recommendation, but also feels compelled to get more information about it and consequently purchase it. All of which is done intuitively, without struggling with the system.

Another fundamental tactic to support recommender system strategies is an efficient costumer database. Customer files should contain relevant information and be properly catalogued so that pertinent recommendations are made by IS. Furthermore, it is important that the algorithms used in the recommender systems are able to accurately access the information in the costumer files.

With the emergence of m-commerce, scalability of recommendation systems is also an important business tactic. Recommender systems algorithms vary according to the device being used (Tveit, 2001). Therefore, in order to fulfill the strategies and goals that predict an increase in costumer base and satisfaction through this type of systems, it is important to have the system standardized in all channels of communication used by consumers use to purchase. Correct application of tactics will collaborate to the accomplishment of business’ strategies and consequently goals and objectives, contributing to a company’s success and competitiveness.

//iii.Objectives//

The results obtained through the application of strategies are measured to determine how effective the recommendation system is in supporting the organization's objectives. First of all, the accuracy and time expansion of displaying query results depend on the algorithms used within the system. On that note, algorithms should be as efficient and effective as possible. Meaning, when algorithms show more precise results, it is effective. When algorithms are faster in producing suggestions, it is efficient. This is relevant in the case of digital television or library catalogs in which prolonged or random research results in user frustration. Especially now-a-days, customers demand integration of their home networks with mobile devices (Zhu, J., et. al. 2008). Meaning, users expect faster results from all appliances that use recommendation systems.

Once algorithms achieve a certain level of efficiency and effectiveness, the system becomes user friendly. End user satisfaction is essential to maintain continuous use of the system and keep individual interest in the suggestions (Felfernig et al, 2008). This is significant in e-commerce because if an organization like eBay can increase customer perusal within their website, the more likely they are to encounter a desired good and eventually purchase products. With more sales, the higher the profit. In summ, the IS measures time spent on queries and stores user trajectory to build user profiles that detect preferences, habits, and market trends (of related goods too).

Furthermore, managers in e-commerce also aim to decrease costs by saving time and circulating inventory. The less time a product is stored in some kind of warehouse, the less costly it is for the company to maintain the good. Thus, relating a variety of products while the consumer shops online can generate more sales of goods that were not going to be purchased at first. In addition, recommender systems may be programmed to pair a star item with a less popular one, so even the products less likely to be sold are.

It is essential to define the target audience and own product and service perceived values to gain competitive advantage (Evans 2010). With this, one of the main objectives of any organization is customer recognition and end-user satisfaction. This can be measured through instances of usage and popularity of the system. Content users attract more individuals and unhappy customers drive away thousands of potential buyers. At the same time, this serves as a selective base for the target audience to be reached out to. This can be applied in e-commerce and other areas that use recommendation systems with their search engines. An example of how viral end-user satisfaction can be is Google. Word of mouth on how researching via Google produced better results, the company ascended and dominated the cybercommunity and in a couple of years became the conglomerate it is today.

//iv. Goals// The core goals of implementing a recommender system are to increase efficiency, effectiveness, customer satisfaction, and ultimately profits. Specific business goals that can be achieved through this type of system in order to support a company’s mission are:

- Increase sales and profits by enabling cross-sales and up-sales through recommender systems. - Increase product awareness through recommendation features that show the users products they wouldn’t have seen otherwise (or sought for). - Improve costumer satisfaction, and therefore retention, by reducing search time and providing an intuitive graphic user interface. - Increase website interactivity by drawing upon customer preferences and expertise to provide recommendations. - Rank customer preferences in order to manage inventory more efficiently.

If a business applies the system properly and meet its objectives, according to its tactics and strategies, it will succeed in achieving its business goals. This success is crucial for a company to obtain competitive advantage in its industry. One example of a company that has, so far, effectively implemented a recommender system to support its business goals is Amazon.com.

Amazon’s mission is to “start with the customer and work backwards” (Amazon.com). This is perfectly coherent with goals that involve recommendation IS. The company’s goal is to add value to the customer’s online shopping experience by lowering search costs (Garfinkle et Al. 2008) and personalizing the online shopping experience through recommendation systems.

The collaborative system saves the users time because it shows them other things they are likely interested in, without having to browse through the thousands of products they have available online. The recommendations are presented in a graphic user interface that is simple, clear and standardized throughout the website. Therefore the overall experience of the user with the system, and consequently with the company, is positive; fulfilling the goal of customer satisfaction. Furthermore, a satisfied costumer is more likely to come back to the website and make more purchases at Amazon.com; fulfilling the goal of customer retention. A satisfied, returning costumer, makes more purchases, which means more profits: the bottom line goal for any company.

Another interesting aspect of Amazon’s system is that it helps improve efficiency of inventory control. Cross-sales proposed through recommendations help clear excess inventory (Grafinkle et Al. 2008). Products that wouldn’t be sold because they weren’t sought for are now effortlessly available to consumers who might be interested in them, thus increasing the likelihood of sale. The system accomplished another goal: increasing efficiency in business processes.

Amazon is an example of a company that has, for the most part, successfully implemented a collaborative recommendation system to support its goals. However, when using collaborative systems to increase costumer’s satisfaction, a company needs to be aware that such systems are strictly based in algorithms and is, therefore, limited. There is more to consumer’s choices than what the algorithms of an IS can calculate (Manousellis, 2008). Personal preferences and culture, for example, can’t be assessed through numbers or descriptions in a customer file. This limitation, if not properly addressed, can be a reason for failure of a recommender system application. Thus, managers that decide to make use of this system need to ensure that these other aspects are covered by the business goals as well, in order to guarantee the company’s success.

//**Conclusion**//

From the research and critical study made in this paper, we conclude that Recommendation systems can be an important asset for decision-making managers. There is even more value added specifically in the e-commerce area. On one hand, costumers benefit from an overall positive shopping experience through personalized services and interactivity of systems. On the other, businesses benefit from the increased sales that result from cross sales and push sales facilitated by recommender systems (which increase the bottom line profits). However, since this type of IS is still relatively new, it should be used with caution. It might have some limitations, which can hamper the company’s progress if implemented inadequately.

Therefore, it is important for decision-making managers to consider the business’ fundamental goals and mission before adopting a recommendation strategy to ensure it will appropriately support the company’s operation and increase both efficiency and effectiveness. For this reason, companies can still be successful despite not possessing state-of-the-art technology. IS concerns the circulation and coordination of information, which can be even further improved and accelerated through IT. Nonetheless, a poorly structured IS combined with IT expedites wrong and faulty information to businesses and customers, leading to unwise decisions.


 * Works Cited**

Best Buy. Best Buy Canada Ltd., 2010. Web. 30 March 2010. 

Evans, Max. Lecture Week 9 – eCommerce. University of Toronto. University of Toronto Mississauga Campus, Mississauga, ON. 30 March 2010. Lecture. Felfernig, A., Teppan, E., & Gula, B. "User acceptance of knowledge-based recommenders". Personalization techniques and recommender systems. Ed. In Uchyigit, Gulden and Ma, Matthew Y. Singapore: World Scientific Publishing Co. Pte. Ltd., 2008. 249. Print.

Orkut Beta. Google, 2009. Web. 30 March 2010. 

Zhu, J., et. al. "Content classification and recommendation techniques for viewing electronic programming guide on a portable device". Personalization techniques and recommender systems. Ed. In Uchyigit, Gulden and Ma, Matthew Y. Singapore: World Scientific Publishing Co. Pte. Ltd., 2008. 223. Print.

Nielsen, Jakob, and Jonathan Levy. "Measuring Usability: Preference Vs. Performance." //Communications of the ACM//. 4th ed. Vol. 37. New York: ACM, 1994. 66-75. Print.

//Amazon.com: Online Shopping for Electronics, Apparel, Computers, Books, DVDs & More//. Web. 30 Mar. 2010. <[|http://www.amazon.com]>. Garfinkle, Robert, Ram Gopal, Bhavik Pathak, and Fang Yin. "Shopbot 2.0: Integrating Recommendations and Promotions with Comparison Shopping." //Elsevier - Decision Support Systems// 46 (2008): 61-69. Print.

Manouselis, Nikos. "Deploying and Evaluating Multiattribute Product Recommendation in E-markets." //International Journal of Management and Decision Making // 9.1 (2008): 43-61. Print.

Tveit, Amund. "Peer-to-peer Based Recommendations for Mobile Commerce." //Proceedings of the 1st International Workshop on Mobile Commerce//. International Workshop on Mobile Commerce, Rome, Italy. New York: ACM, 2001. 26-29. Print.

McNee, Sean M., Shyong K. Lang, Catherine Guetzlaff, Joseph A. Konstan, and John Riedl. "Confidence Displays and Training in Recommender Systems." //Human-Computer Interaction: INTERACT '03//. IOS, 2003. 176-83. Print