Combosquatting

Methodology

We initiated this study by selecting a group of 30 brands, based on overall brand strength and the number of search terms regularly associated with them. The brands represent a wide cross section of products and services in order to best inform our analysis of combosquatting practices across multiple industries. Once we selected the brands on which we would focus, we harvested a list of the top 50 search terms that included those brands using a search engine keyword suggestion tool; this was done in order to determine what phrases Internet search engine users are looking for when they search for these brands. After identifying the search terms, we used Wordtracker™ to check the daily searches across three major search engines—Google, Yahoo! and MSN—and combined these daily averages into a monthly average for each search term. We then used the search terms to create a list of 1,500 domain names. For example, the search term “Craigslist Washington” became “craigslistwashington.com.”

Figure 1 [+] : This figure contains examples of brand+keyword domain names that may be combosquatted or are owned by the appropriate company.

FairWinds’ traffic estimation tool was used to determine the monthly traffic for each domain name. The traffic for each individual domain name was then compared to the search counts for their corresponding search terms to create a ratio between the two. This ratio demonstrates the relationship between keywords associated with brand names that are entered into search engines and the same keywords that are entered directly into a browser as a domain name.

In order to better analyze and understand the data, we removed any combinations of search terms that did not have reliable traffic data. We also removed any domains that received a high amount of traffic as a result of being communicated in advertising—for example, thefacebook.com used to be the official and sole working domain name associated with Facebook. We found this to be logical due to the increased traffic received by communicated domain names that may not be intuitive to Direct Navigators for any other reason. Because statistical measures used to eliminate outliers are not able to distinguish between removing communicated domain names and removing true unexplained outliers, we expelled the five hundred domains with the highest and the five hundred domains with the lowest ratios between traffic and search, and retained and analyzed the middle 500 combinations. Choosing this data set helped ensure that the domains in our study were entered by users who were guided solely by their intuition.

We looked at the distribution of the ratios for the 500 domains qualified to be in this study hoping that the relative strength of this relationship could facilitate a better understanding of both the potential threat from combosquatters and the potential benefit for brand owners by showing just how many Internet users type in these keywords expecting relevant brand content.

Key Results

• 50.6 percent of sites tested contained PPC advertising
• 22 percent of sites tested were legitimately used for brand purposes
• 29 percent of brand+keyword combinations used in this study received more Direct Navigation traffic than search engine hits
• Of those combinations that received more type-in traffic than search engine hits, 59 percent were not owned by the intended brand owner
• 75 percent of brand+keyword combinations that were not owned by the intended brand pointed to PPC sites
  

< previous page | next page >