t is important to increase loyalty of the visitors to the web site. This would be manifest as more frequent visits by each user (Time spent by each user cannot be considered – personalization may actually reduce the time spent by each user by quickly giving the user what he desires, rather than the user himself looking for it.) The following statistics need to be constantly monitored:Number of unique visitorsFrequency of repeat visitorsConversion ratios:-Number of visitors becoming account holders-Number of account holders becoming customers-Number of customers making repeat purchasesThese could be monitored for all visitors or for segments of visitors, as well as for various features of personalization. Each of these statistics should be subject to comparison with a control sample. This control sample should be from the historical data collected by the insurance company, when personalization had not been implemented. Comparison could also be made with the corresponding data for a competitor’s web site, if possible, especially one that is faced with the same external environment as the insurance company. This would also help to establish benchmarks for the performance of the site.12.1Evaluating the contentThe system needs to constantly analyze whether the content that had been put on the site is actually fulfilling its purpose. For this, it is important to take feedback from the users. It is proposed that the articles on the web site be followed by the following questions:How relevant did you find the article? (Very relevantsomewhat relevanttotally irrelevant)How detailed did you find the article? (Too detailed and complex adequate detailstoo sketchy, more information needed)The first question finds out the efficiency of the information, while the second refers to specificity. Such an analysis would help to fine tune the personalization effort. However, it is proposed that not all users be asked for feedback. Instead, wi...
