Problem: Google notoriously collects and leverages an immense amount of user data to better improve ad targeting, click through rates, and ultimately their bottom line. These algorithms are largely a black box, meaning we as users don’t have insight into how our personal data is being leveraged within them. Unskew is an attempt to identify how our search results are being impacted by unique data associated with our accounts.
How it works: When a Google search is run with the extension installed, Unskew takes the search term and sends it to our servers. From there, we use a network of proxies to run the same search from a user agent with no connection to you or your account. We scrape these unaffiliated results and return them to your browser where the extension compares them to the results you yourself received.
Hypothesis: By controlling for the variable of user identifying information and removing it from the equation, disparities in search results represent attempts by the recommendation algorithm to personally tailor results. Identifying the extent to which this occurs is vitally important, given Google’s near monopoly on the search space. Potential negative impacts could involve users increasingly becoming encased in an echo chamber of their own making in order to maximize click through rates, or could indicate the willingness to engage in even wider scale manipulation of search results across user accounts which are more difficult to detect.
Feedback: We hope to hear and engage with feedback regarding our methodology, UX, and how this tech could be better leveraged to benefit users and the UX, and answer any questions. Outside of our immediate peer network this is our first attempt at gaining wider feedback.