A huge number of brands rely on Google search ads to drive in-store traffic. However, some brands question the accuracy of the Google Ads attribution.

Google uses AOV (“average order value”)  ―a plug variable that the marketer provides. Some consider this measure inferior to pure online tracking of an online sale, which measures actual AOV per customer. More on this below.

Google has published its methodology, and a large scale study including, roughly 5MM people, suggesting in-store visits are more-or-less effective.  We’ve done our own research, granted at a much smaller scale, to obtain our own first-hand data and experience.  with the efficacy of using Google to track in store visits. Our experience did in fact align closely with Google’s study. 

For the most part it worked surprisingly well  ―one exception was when you visited and had material dwell time between two attached retail stores that share a wall between them.


Does Google Track a User Who’s Driving Past The Store As a Sale?

We’ve heard retail brands, first hand, suggest that online-to-offline sales attribution doesn’t work. The one reason they’ve cited is because they feel or believe that a user who viewed online, and subsequently drove or walked past the store near them will be attributed as a sale.

Based on our experience and studies across brands, this isn’t true for Google Ads/ Google Maps. The location accuracy is factually better than that, and the dwell time requirements screen out the vast majority of “drive by’s”

This concern likely surfaced when cellular companies shared data years ago, to produce triangulation data to track user location offline. This is possible, but the accuracy is questionable –especially when the cell towers are miles apart. They use signal strength between multiple towers to calculate your estimated distance between each of them. This can work, but is less accurate to “fine GPS” location, which the Google Maps app and Waze, which was acquired by Google in 2013.

So it’s important to consider these are very different approaches using very different data sets. We are confident that Google’s tracking and attribution is substantially more accurate than traditional cell tower triangulation. Though it is possible in the future that 5G carrier technology, primarily 5G cell towers being smaller and more widely distributed, could work more effectively.


How Does Google Attribution and Technology Work?

Leveraging Google Maps app and location data, Google knows the exact coordinates and borders of millions of businesses globally.

So logically, Google integrated Google Maps to match location history for hundreds of millions of Google users with Google Maps data across roughly 2.05 million physical retail businesses  ―and growing.

Google publicly shares that they integrate a large number of signals in order to most accurately measure visits. While we may not know all of them, these are some you’ll find if you do the research:

      • Location History
            • If users have enabled location tracking on their mobile devices, Google can use this information to determine if they have visited a physical store after clicking on an online ad.

        • GPS location signals, including fine-location (accurate to 1 meter)

        • Google Maps Street View data
              • Google uses signals from Wi-Fi and Bluetooth beacons in physical stores to determine if a user has entered the location.

          • Street View coordinates of hundreds of millions of stores globally

          • Google Earth

          • Wi-Fi signal strength from specific store locations.
                • Google uses signals from Wi-Fi and Bluetooth beacons in physical stores to determine if a user has entered the location.

            • Actual user searches

            • Duration of visits both online and offline

            • Proprietary 1 million Google user cohort opted into sharing on-ground location history

            • Google proprietary modeling

            • Surveys
                  • Google may also use surveys to gather information directly from users about their shopping behavior, which can help to confirm store visits.


            How Does Google Ensure Accuracy of Offline Attribution?

            Google advises they have also surveyed over 5 million users who have visited a store, validating first hand at the individual level that they did in fact visit. They used this feedback to improve their matching and claim their in-store attribution is now “99 percent” accurate.

            While 99% is not a great value for systems uptime or national security applications  ―it is a very compelling value for online-to-offline purchase attribution.

            Our experience also shows that when properly configuring campaigns, and using in-store conversion, we do see a strong correlation alignment between spend and efficacy. We have also tested across categories and brands to measure how well the correlations hold up. The one exception is when you’re testing Offline Conversions for the first time with a small budget —or your a small business that doesn’t get enough visits for it to impact your store purchase volumes. The incremental sales can more easily ‘wash out’ in weather or time of day, day of week, time of year or even local due to local traffic issues. Larger budgets generally shows a more clear lift sooner, as the behavior isn’t ‘washed out’ in a short term trend, like the aforementioned. 


            Your Role in Google Offline Attribution Accuracy & Value

            There are two major vehicles in using Google Ads and Offline Attribution that you affect.

                1. Average Order Value

                  You must provide the average order value to Google for all offline sales Google reports. For brands who can’t quite get comfortable with Google’s offline attribution, a simple technique is to assign your own confidence level  ―for example, if you think it over-reports by 10%  ―you can multiply your standard in store AOV by 0.1 and you’ll curve your sales values down. You could use this as a buffer too.

                  Bear in mind there are some businesses that have considered curving them UP, as they find that users who have material online engagement prior to an in-store visit actually spend more  ―especially if those ads were targeted optimally to the right customer.

                1. Upload Your Offline Transactions to Google

              You can improve AOV accuracy in online-to-store attribution by uploading your offline transactions to Google. You’ll also have to include the Google Click ID (GCLID). With this data, Google will stitch together the click (from search), the visit to the store (with Google Maps/ find GPS data) and transaction, including time/date stamp, where Google’s machine learning will match transactions based on transaction time, search data and dwell time to a specific sale that happened in the store. This works with up to 90 days of offline transaction data.



              Google’s store visit tracking can be accurate, even if we assume there are some limitations to the technology. For example, if a user has disabled location tracking or has their mobile device in airplane mode, Google will not be able to track their physical location. 

              Additionally, Google’s measurements may not be perfect, and there may be some margin of error, even if modest. Concurrently, Google has a reputation for continually investing in improving its technology, and the accuracy of store visit conversion attribution is generally considered to be high.

              For those who still skeptical of the revenues Google offline attribution shows as a result of a YouTube video view or a ad click through to a website, you can simply apply a ‘confidence factor’  ―multiply your average store AOV, by a value of less than one, e.g., “0.9” to reduce the value of attributed sales by 10%, and recognize a level of store revenues you believe to be most likely.