Google Has A Search Bubble?

Google has always had the spotlight when it comes to search since it revolutionized the way users access the web. It’s grown to a point where in the last year they consolidated all of the privacy clauses into one, giant blanket one that affected all of their online properties.

It wasn’t a major change in the way you use any property that Google owns, as for the most part the privacy policies contained the same verbiage, but a recent ruling in the EU may cause everything to be backtracked. It was decided that because Google didn’t offer an out clause on the new blanket privacy policy, that it breaks the use privacy laws in the EU. If the decision stands, then the door may be open for every other country to take up arms and work at forcing the company to revert back to it’s old way of handling privacy policy. Not that it would really affect anything, all the change to a single policy did was serve to streamline usage of Google websites and properties. Change is always a good thing online, however taking a step backwards is typically viewed as the wrong direction.

An example of moving forward with search, I’ve mentioned a handful of times in the blog, is the DuckDuckGo search engine. Recently the small search company produced a video where they talked about how Google has each use caught in what they called a search bubble. Where they took more than 100 users, ensured they were not signed into their Google accounts and had them conduct searches on specific terms and captured their results.

What they found, was that even when the users were not signed into their accounts, and even in the same geographical area, they received differing results pages. It’s not a revelatory video really, as Google isn’t the only company on the web that utilizes browser cookies to determine who a user is and what they may like. Not to discredit what DuckDuckGo is hinting at, but with such a small sampling, and by allowing users to use their personal computers without clearing any session cookies, it’s no wonder the results were different for each user. Perhaps with the addition of a control group, a group of 20 users or so who were using completely clean installs of a browser and OS would help balance their results.