Mountain View has made technologies available for the so-called under Open Source license Differential Privacy, these are essentially the tools used by the developers of Google to extract relevant information from the data (technically insights) without the results obtained being associated with the users who supplied them.
Libraries for Differential Privacy can be exploited freely by anyone who wants to have detailed statistics with which to improve the quality of their service (just think of any store for e-commerce) while having the guarantee that these will not be collected or treated in violation of privacy.
The Differential Privacy is based on a mechanism whereby the data is partially altered so that it cannot be traced back to the person who sold it, it will be possible to process all the available information obtaining flows and trends from them, but in no way will it be possible to profile users from the data that would allow them to be identified.
To clarify the usefulness of this approach, the spokespersons of Big G wanted to refer to a particularly sensitive sector, that of clinical data. The Differential Privacy could be used, for example, to calculate the time spent by patients during hospitalization periods, without being able to associate a specific stay with a specific person.
As pointed out by Miguel Guevara, Product Manager of Google, the Differential Privacy can be just as useful in a business environment, for Public Administrations or for Scientific research, all areas in which confidentiality should be respected without the protection of users compromising the quality of the processed insights.