Google provided advertisers with an update on 11/27/2019. The issue has been completely fixed, and all reports are accurate now.
While the issue has been resolved, the bug led to a false 53% increase in conversions across our clients for the affected date range. Check out how we analyzed this issue below!
What you’re seeing is a screenshot of Google Ads performance changes across a handful of clients between 10/31/19 – 11/9/19 and 11/11/19 – 11/20/19. The remainder of Seer clients saw trends pretty similar to those depicted here.
As you can see, we’ve done it. We’ve finally cracked the code to breakthrough performance on Google Ads.
Ok, maybe not quite…
Google released an update this past Thursday, November 21st, notifying advertisers that a reporting bug is causing Google Ads platforms to over-report on conversions across search and shopping campaigns. The data affected by the reporting error spans from 11/11 to 11/20. One client, for example, saw an additional 865 conversions during this period.
Details Behind the Reporting Bug
- According to Google, this bug affected advertisers that are leveraging ‘non-last click’ attribution. (Which is a majority of advertisers)
- The bug affected data in reports coming from the Google Ads interface (Reporting Tab and UI Exports), as well as any downloads coming from the Google Ads API, and Google Ads scripts.
- Although conversion volume was inaccurate, this data was not supposed to be affecting the machine learning models behind automated bidding strategies according to Google.
- Google has corrected the issue as of November 27th.
Advertisers often base the success of their paid media efforts on how many conversions they generate, the cost of those conversions, and the revenue that they generate. For those more familiar with the industry, we’re used to looking at KPIs like CPA and ROAS when determining the level of success for our paid search and shopping campaigns.
As you can see in the screenshot from some of Seer’s clients, an artificial increase in conversions can significantly impact CPA, which makes these campaigns appear much more efficient and profitable than they really are.
This can lead to misguided shifts in strategy, especially ahead of the big Black Friday/Cyber Monday weekend.
Conversion data impacts how advertisers adjust bidding strategies, qualify search terms as valuable or irrelevant and allocate budget across campaigns based on performance. With inaccurate conversion data, digital marketers’ hands were tied when trying to make any of the aforementioned optimizations based on data from 11/11-11/20.
For example, all Seer clients saw conversions increase an average of +53% for the date range affected by the bug. This issue may have skewed projections, and potentially jeopardized the success of any Black Friday efforts advertisers invested time into.
Here’s how we looked across 20 days of client data in 5 minutes to come up with a solution for each client affected:
Most Google Ads accounts received an alert message once the bug was resolved. If you still want to check your data, follow these steps.
- Check for multiple conversions coming from single clicks in search query reports.
- Look for spikes in the Google Ads UI around 11/11 and 11/12.
Adjust your reporting for the time frame of 11/11 – 11/20. You can do this by discounting conversion volume using historical performance from recent weeks.
💡 Pro Tip: Apply your historical CVR to your click volume from 11/11-11/20, and this should provide you with more accurate conversion numbers for November as a whole.
Want to be the first to know what anomalies we find in our data when big shifts happen in digital marketing? Sign-up for our newsletter below: