Sidebars & Backgrounders
Analysis of Nielsen, Hitwise and comScore Rating Services
Measuring traffic on the Internet has become increasingly complicated. A number of methods have emerged, but no single methodology has been accepted as a standard.
Nielsen and comScore use a panel of Internet users to estimate total U.S. Internet traffic. Just as a telephone polls contact a sample of Americans, Nielsen and comScore contact a sample of Internet users who agree to share how they spend their time on the Web. Internet users who participate then download software on their computers that tracks their online visits without attaching any personally identifiable information to the traffic data to ensure anonymity.
Nielsen uses a random sample of Internet users by collecting its panel with telephone calls, the method used by most pollsters today. Achieving a random sample of participants for the panel is a key advantage. With a random sample, Nielsen can take the traffic data and more effectively generalize to U.S. population of Internet users as a whole. Nielsen runs two panels – one consisting of people at work and the other of people at home – for a combined total of 30,000 participants.
In 2008, Nielsen began testing a product that tracks how television viewing and Web browsing interrelate. It combines existing Nielsen methods for measuring television viewership with Nielsen Online’s sample-based system for estimating Internet usage. During the year, NBC tested a system called Total Audience Measurement Index that was designed to measure and analyze traffic and viewership for the network’s Olympics broadcasts and webcasts.
ComScore recruits what it calls a convenience sample instead of a simple random sample by offering incentives to participants. ComScore then applies statistical methods to adjust, or weight, the results to reflect the demographics of the actual online population.
For example, after it obtains traffic data from a panel, comScore analysts may discover they have a smaller percentage of males than the online population at large. They then add more results from males so that they are correctly represented. ComScore says this gives it an advantage because it uses more people (about 150,000) and maintains three panels – at work, at home and at universities — to ensure that the data capture how students are using the Internet differently from adults at home or at work.
Hitwise takes a wholly different approach. It does not gather data directly from individual computers as comScore and Nielsen do. Instead, it gets the data from Internet service providers (ISPs) who aggregate traffic data across all the individuals to whom they deliver Internet access. Hitwise provides ISPs with proprietary software that allows them to analyze website usage logs created on their networks. To ensure the data are representative, Hitwise says it collects “from a geographically diverse range of ISP networks in metropolitan and regional areas, representing all types of Internet usage including home, work, educational and public access.” The sample of ISPs, however, is not a purely random one.
Hitwise feels its data offer some advantages. First, it reports data using specific website addresses (like www.CNN.com). ComScore and Nielsen report data that include all related websites (like money.CNN.com), a definition that can change over time. Hitwise data also put a premium on anonymity. Because it collects aggregate data, it never has access to the personal information of Internet users.
Different methodologies often result in different results, and that is the case here. Nielsen and comScore, for example, both rank Yahoo News, MSNBC.com, CNN.com, and AOL News as the four most-visited sites. Both rank AOL News as fourth. Hitwise, however, ranks AOL News as the 11th-most popular news site.
While Hitwise ranks the Drudge Report as the fifth-most popular site, it does not register on the top 25 sites as listed by comScore or Nielsen.
Hitwise included a number of sites that are not generally regarded as traditional news sites and did not appear on the Nielsen or comScore lists. These include Yahoo Weather, Yahoo Local, and TVGuide.com. Both comScore and Hitwise listed some that Nielsen did not list, including the Weather Channel and Weather Underground. Nielsen did not include any of those websites in its categorization. According to Hitwise’s press office, differences materialize largely because of what the company categorizes as news.