A big problem facing by Bluefin Labs is that Bluefin’s data reliability. First, Bluefin’s data mainly come from Twitter, a real-time communications platform with 200 million monthly active users globally. From what CBS Television has commented, 70% of Bluefin’s data come from Twitter, and only 10% of CBS’s market is on Twitter. In a way that Bluefin could not provide what client really needs. Also, Bluefin’s reliance on the data from Twitter may cause a selection on biased sample of audiences since the majority of Twitter’s users tend to be young adults. Furthermore, Bluefin collects its demographics by retrieving users’ public screen names. Among Twitter’s 200 million monthly active users, many of them are fake users and maintain duplicate accounts. And there could be groups of spammers who are paid working together to influence reviews. It is hard for Bluefin to segment and target different groups of users. Hence Bluefin’s data source are not actually representative for all groups of users. Nielson, the television ratings girant, is a robust competitor to Bluefin Labs. For representativeness, CBS Television supplemented its Nelson People Meter data with a proprietary viewer panel. If it wanted viewer reactions to a show, or wanted to know reactions from a subset of the penel with a propensity to influence other viewers, it had only to email its question to the panel and receive, within three or four days. However, Nielson’s sample is not random in the statistical sense. A small fraction of the population is selected and only those that actually accept are used as the sample size. To meet the needs of clients and gain a competitive advange, Bluefin Labs will have to provide much larger samples and diversified data source. Integration of several social platforms and networks could increase sample size and select a more unbiased sample of audiences. Bluefin could also develop an improved algorithm to detect fake reviews and groups of spammers. Checking the completeness of personal profile and user login frequency is a way to exclude fake users. And timing will be helpful when spotting spammers. Often spammers post similar reviews within a short time of each other and use very similar language. With an developed algorithm, Bluefin Labs could prove its data reliability.