fake search.live.com referrals - quality testing or spam?

Are the fake referrals from search.live.com part of a quality testing regime or is there a classic spam motive?

It is theoretically possible for search engines to test for cloaking, (where a website deliberately serves search engines with different content from ordinary visitors), by disguising their robots as ordinary visitors, and the semi-official claim by Microsoft that their fake referrals are for quality purposes is almost viable.

Sending requests with keywords in the HTTP referrer header seems unnecessary though, and the boffins at live.com must know that their activities will be corrupting traffic reports and making it look as though live.com is generating real traffic. Because of this it is difficult not to assume that the exercise is partly a deliberate attempt to improve live.com's profile.

IP Ignoring

Microsoft have admitted to sending these fake referrals and have reportedly suggested that the activity "should not be statistically significant for any sized website". So that's OK then? I don't think so.

There have been warnings that blocking the referrals could prevent websites from being indexed by live.com so I've added a new component to MJB Data's tracking system call 'IP Ignoring'. Whereas 'IP Blocking' prevents specific IP blocks from accessing content, 'IP Ignoring' serves the content normally but flags each page request as fake.

While developing and testing 'IP Ignoring', search.live.com has generally disappeared from the list of top 10 referring domains. It could be argued that the total numbers are relatively insignificant, but it's likely that live.com is appearing on millions of website traffic reports that it otherwise wouldn't be.