When Google rolled out its new recommendations for Search Console, one notable highlight was the suggestion to bulk export your data because “Search Console only displays 35% of your performance data.” While this isn’t entirely new information—Google has long anonymized a significant amount of data in Search Console—this recent clarification sheds light on the extent of that anonymization.

In July 2022, we noted that, on average, around 50% of queries in Search Console were hidden from view. Google didn’t dispute this; instead, they updated their documentation to elaborate on the limits and anonymization practices surrounding this data.

Interestingly, many users seem to be unaware of this issue, especially now that it’s prominently featured in the new recommendations section (though it’s worth noting that many users may not even see this section).

Understanding the 35% Data Limitation

John Mueller from Google addressed the 35% figure in a comment on LinkedIn, stating, “Most sites see all or most of their data.” This suggests that larger websites might experience greater data limitations. He elaborated that the number varies depending on the individual site, adding that the recommendation to bulk export appears when Google detects a significant discrepancy in the data.

Mueller further clarified, “Even if not all data is displayed individually, the overall graph remains accurate. Utilizing the API can provide more data, so tools like Search Analytics for Sheets can help users get a better picture. It’s important to focus on the actionable data you need—while there are options to dive deeper, most users probably don’t require that level of detail.”

Moreover, that 35% figure can fluctuate for different sites. As mentioned in previous studies, it was closer to 50% for many sites tracked by Ahrefs back in 2022. “It really depends on the site—there’s no fixed percentage,” Mueller emphasized.

So, while this revelation might seem new to some, the takeaway is that if you want a complete view of your data, exporting it has always been advisable. If you encounter this recommendation, it may be a signal to export your data. Conversely, if you don’t see the recommendation, you might still want to consider exporting your data and comparing it to understand the limitations better.

For further insights, check out the forum discussion on LinkedIn.

Update: Following my post, John Mueller commented on Mastodon, stating: “Yes, not new. The recommendations serve to inform users about potential process adjustments, especially for anything they might have overlooked. Not everything will be actionable, and not everyone has the time to implement every suggestion. Perhaps we should label it as ICYMI (In Case You Missed It)!”