Do Google and Microsoft use their Gmail and Hotmail/Outlook free email services to look for new websites to add to their search engines?
An experiment to test whether GMail and Hotmail/Outlook are data-mining their users' emails to feed the Google and Bing search engines.
Alphabet has the Google search engine and the Gmail email offering. Microsoft has the Bing search engine and the Hotmail (aka Outlook) free email offering. If I ran a search engine, I'd be looking to improve the service in any way I could, and that might include searching free services I provided elsewhere to look for new web addresses to add to my search engine.
Google already scans emails for words that can be passed to their advertising engine to enable it to zero-in on your interests, so why wouldn't they use their email facility to provide the much more benign service of looking for new web addresses? Nobody really wants advertising, but people who run websites want their sites added to search engine indices, and the people emailing their links around presumeably wouldn't be too bothered either way (especially if they're already okay with the idea of Google scanning their emails for advertising purposes).
So... do Google and Microsoft do it? The creation of Changewatching presented a good opportunity to test their systems. Google and Bing both provide online forms to allow new websites to be submitted for indexing. Normally I would use these when creating a new website, but this time around, I decided to test the internet's big names' free services.
As part of testing Changewatching's email interface, I had to try sending emails from various popular email providers, just in case there were any quirks in their email handling I hadn't thought of.
First under the microscope was Microsoft's free online email offering, Hotmail (now rebranded as an online version of 'Outlook'). Despite presumably never having dealt with any mail from the Changewatching.co.uk domain, Hotmail flagged Changewatching's email as spam. Boo!
Over a month after sending various email to and from the Changewatching domain, there is no sign of Changewatching turning up in Bing's search results.
10 days after a similar test with Gmail, there is no sign of Changewatching in their results, either. Both tests were conducted using free email accounts that I locked down to the most private settings available. 10 days really isn't long enough to test Google but I am tired of this site languishing in the internet purgatory of being online but not indexed, so I am bringing this test to a halt.
Note that Google and Bing's activities were tested only on the basis of whether or not Changewatching entered their search engine indices. Normally I would also check the web server logs for evidence of their web crawlers' visits, but as Changewatching is set up for maximum possible user privacy, its web logs are excluded from the tracking/analysis system I use for other sites.
Conclusion
Microsoft's Bing / Hotmail / Outlook: Negative. Based on a number of tests conducted over a month, it appears that Microsoft does not trawl through their user's emails to find new search results for their Bing search engine - at least not from a privacy-locked account, on this occasion.
Alphabet's Google / Gmail: Inconclusive. Based on testing conducted so far, Google does not appear to trawl emails for search results - at least not from a privacy-locked account, on this occasion. However, the testing was suspended after only 10 days. Previous experience of monitoring the speed at which changes propagate through Google suggests that this is probably not a long enough test and therefore the result is inconclusive.
Further Research Required
This represents only one data point, and the result for Google is inconclusive. To get a better picture of whether emails are scanned for the benefit of search engines will require many more data points, and therefore many more tests of this sort conducted over a long period of time. It also requires testing from email accounts that have not been privacy-locked. There are also many variables that might affect the result - for example, Gmail might 'trust' older email accounts more than newer ones (in much the same way that older websites tend to get better search rankings than newer ones). The policies and practices of Microsoft and Google may change over time.
As I bring more websites online in future, I will likely conduct further similar tests.
Link to this text