comScore has developed a statistical methodology to ensure that its network is a reliable mechanism for projecting findings to the Internet population. This enables comScore’s clients to gain a truly accurate picture of Internet behavior and purchasing patterns.
The comScore Global Network is the largest source of cross-site customer information ever assembled. This vast network of members delivers the highest quality view of customer surfing and buying activity across the entire Internet.
To ensure a well-balanced measurement audience, comScore understands the importance of a multi-channel recruiting effort. We use various online recruiting mechanisms, as well as traditional random direct mail and random digit phone dialing to add new members.
We are currently adding thousands of new members a day to our network by offering a compelling value proposition to entice consumers to join:
- Doubling of download speed
- Attractive sweepstake prizes
- Opportunity to impact and improve the Internet
- Total anonymity of personal information
- No targeted ads or intrusion
- Minimal membership requirements
All Segments of the Internet Population Represented
All demographic segments of the Internet population are represented in the comScore Global Network and each segment includes large numbers of members. For example, our network includes tens of thousands of high-income Internet users — one of the most desirable and influential groups to track, yet the most difficult to recruit. All told, we are measuring Internet activity and accumulating online purchasing data from more than one million individuals.
Balancing and Projection
To project the purchasing and surfing of Network members to all Internet users, the members are divided into a large number of groups based on their demographics, such as income, and on characteristics of their Internet use, such as where it occurs (from home or work) or extent of online use. Projection weights are then calculated group by group, so that at the same time that comScore projects purchasing and surfing of Network members to Internet users who are not in the Network, it eliminates the effects of the over — or under — representation of any group in the Network.
comScore estimates the number and characteristics of all Internet users, which is the population to which it projects measurements, with a weekly survey of 1,000 adults. Random digit dialing is used to select the samples of adults who are interviewed. The respondents are asked whether they use the Internet and, if they do, for information about their Internet use. They are also asked for information that describes themselves and their households. It is the information from this survey that enables comScore to accurately estimate the number of Internet users, which is critically important for producing reliable sales and surfing measurements.