IX. Analysis of Clickstream Data

Report
This topic continued the examination of online market researched, focusing on the collection of B2C data on consumers, products, etc. Sources of data for collection can include internal data (e.g. sales, payroll), external data (e.g. government / industry / competitor reports), and clickstream data. Clickstream data may be gathered using software packages that run under a website, tracking where customers have come from and where they are going, and trailing their activities within the site. These data consist of records of a user’s browsing patterns, i.e. every website and page of every website the user visited, how long the user remained on a page or site, in what order the pages were visited, etc.

Clickstream data allow a business to profile a great number of consumers, which is something that cannot be done in a physical outlet. The data reveal information such as what goods a customer looked at, what goods the customer purchased, what the customer examined but did not purchase, what items the customer bought in conjunction with other items, and what items the customer looked at in conjunction with other items but did not purchase.

Subsequent analysis of profiles can then reveal which ads and promotions were effective and which were not, which ads generated attention but few sales, whether certain products are too hard to find and/or too expensive, etc.

Information that cannot be gathered online is what items are of interest to consumers.

A video entitled “The World According to Google” followed, describing how Larry Page and Sergey Brin invented the software that revolutionized how people search for information. Google’s relevance ranking and its ability to allow people to search for language in its everyday usage resulted in a public, useable search engine unlike prior designs. In 2000, the successful search engine was converted into a successful business by having companies pay per click on their sites. Other features include Google Earth and the program’s inclusion in the past year of business listings, seasonal features and Google books. Google purchased YouTube in 2006 and is considering buying Facebook, through which it hopes to find another way to make money.

Issues that Google face are related to clickstream data, i.e. the ethics of collecting and storing information on user searches. Is this an infringement of privacy? Could this information be used in a court of law? Click fraud, in which people click on a competitor site just to make the competitor pay, is another issue faced by Google, and one with which the company has managed to deal.

Other issues encountered by Google are the effects on books and on research, and the consequences of countries (such as China) imposing limitations on searches, impeding Google’s efforts to provide freedom of information.

Reflection
I can see how clickstream analysis can contribute to market research, though it must be an enormous task to arrange this kind of data into useable forms. E-Commerce 2006 discusses this difficulty as one of the limitations of online market research (p.157). The data needs to be organized, edited, condensed and summarized, all of which may be expensive and time-consuming. Data mining and data warehousing, however, are ways of automating this process.

According to the article “Data Mining: What is Data Mining?” http://www.anderson.ucla.edu/faculty/jason.frand/teacher/technologies/palace/datamining.htm, data mining software allows users to analyze data in order to discover correlations or patterns. Retail, financial, communication and marketing organizations use this process to establish relationships “among ‘internal’ factors such as price, product positioning, or staff skills, and ‘external’ factors such as economic indicators, competition, and customer demographics.” From there, the organizations can assess the impact on customer satisfaction, sales and company profits.

The article also gives some interesting examples of the use of data mining. One is of a Midwest grocery chain that used Oracle software with data mining capacity to analyze buying patterns. Midwest discovered that men tended to buy beer at the same time that they bought diapers on Thursdays and Saturdays. Analyzing further, it was determined that these shoppers did their grocery shopping on Saturdays, while they bought only a few items, like diapers and beer, on Thursdays, to have them available for the weekend. Using this information, Midwest decided to move beer closer to the diaper display and to charge full prices for beer and diapers on Thursdays. The insight gleaned from data mining helped Midwest increase its revenues.