In a blog post announcing Google Attribution, a new AI-powered marketing tool to help clients measure the impact of cross-channel marketing campaigns, the tech giant revealed that it has tracked more than five billion in-store purchases over the past few years.

Google announced Google Attribution in advance of the annual Google Marketing Next event in San Francisco. Google Attribution is a new tool that integrates AdWords, Google Analytics and DoubleClick Search to provide a complete picture of cross-device, and cross-channel marketing campaign effectiveness. The tool assembles data from AdWords, Google Analytics and DoubleClick, then runs the data through an attribution model to analyze the effectiveness of a specific marketing campaign. Attribution then reports the data back to the client in an easy-to-manage format, so that the client can review results and decide on an appropriate course of action with more complete cross-channel data.

Google will roll out an additional feature called ‘store sales measurement’ at both the device and campaign levels in the coming months. Using customer email information at the point of sale, collected for store loyalty programs, merchants can import specific in-store transaction data directly to Google AdWords. This information can then be correlated with a customer’s exposure to digital advertising.

If email data isn’t collected at the point of sale, merchants can use debit and credit card information, which is captured through Google’s third-party partnerships, to analyze in-store purchase data.

The new store sales measurement feature is offered in addition to store visit conversion, which was introduced three years ago. Store visit conversion data is based on anonymized statistics, analyzed from aggregated purchase data that cannot be tied to individual customers. Store visit data consists of modeled numbers based on people who click on advertisements and later visit that merchant’s store.

It is through store visit conversions that Google claims to have analyzed over five billion on-site, in-person transactions to provide merchants with marketing effectiveness information. This provides marketers with solid information on the real-world effectiveness of digital advertising campaigns. Without it, should a customer respond to an ad with an in-store visit rather than an online purchase, the merchant might underestimate the positive effect of digital advertising.

Google also intends to apply advances in deep learning models and machine learning to expand services to larger data sets.

While the announcement was followed up by a message from Google stating that purchase data is tracked in a “secure and privacy-safe” manner, some consumers are concerned that the data could someday be used for a different purpose, or that it could become a target for hackers.

Google users do have the option to limit the company’s ability to track their data, and to control the types of advertisements that they are shown.