Hitwise Stats May Not Be Accurate
By Anna Johnson on March 30th, 2009We quote a lot of statistics in Kikabink News, but it’s worth remembering the old adage: there’s lies, damned lies and statistics (and interpretations of statistics). In Hitwise’s case, it appears the company’s latest search data may not be truly reflective of the market.
As reported by TechCrunch, last week Hitwise said the most searched term on the Internet was ‘Craigslist’. This contrasts with what TechCrunch calls ‘real’ data from Google Trends which shows that ‘Facebook’ is the most queried term, followed by MySpace and then Craigslist.
In Hitwise’s defence, it aims to report on search term popularity across all search queries, not just those made using Google. Google Trends also purports to measure trends across all searches (not just Google searches) but its data is likely to be Google-centric (since Google presumably KNOWS what terms are used on its search engine!).
Hitwise bases its estimates on the usage logs submitted anonymously by Internet Service Providers (ISPs). The only problem with this approach is that not all ISPs – and not necessarily the most popular or representative ISPs – necessarily submit their usage data to Hitwise.
Hitwise claims its data to be representative, but it’s still based on a sample, which may well explain the deviation from the Google data.
That’s why it’s often advisable to look at the data from Hitwise, Google Trends, Compete, Quantcast, and the like, and take note of the TRENDS in that data, more than the actual levels.
Source: Michael Arrington, “New Hitwise Stats Show How Bad Hitwise Data Is,” TechCrunch, March 20, 2009
Related Posts
- Compete, comScore, Hitwise, Nielsen, Quantcast: Whose Internet Traffic Data Can You Believe?
- Hitwise: Google Gets Nearly 73 Percent of All U.S. Searches
- Google Insights For Search – Powerful and Free
- Free Tool: Google Trends To Measure Trends In Traffic and Searches
- Google Launches Google Ad Planner
