Local Memo: GBP Keyword Impressions Undercounted for June, July, LLMs Show Recency Bias, New Guidance on Review Extortion Scams

GBP Keyword Impressions Likely Undercounted for June & July

The News: SOCi, along with others in the industry, observed a widespread drop in Search Impressions from June 24 through July 18 due to a bug in Google reporting.

Though Google confirmed no other insight metrics were affected by this bug, SOCi has observed that GBP Keyword Insights may have also been undercounted at a rate very similar to search—a sign that keyword reporting is directly tied to impressions reporting.

What This Means: Google recently confirmed with SOCi they have no plans to backfill the search impression data. This is disappointing news for those tracking unbranded keyword success, as Google’s Keyword Insights reporting is the best way (besides rank reporting) to show success in unbranded keyword growth. Without these insights, it will be much more difficult to provide data supporting non-branded GBP keyword growth over the summer months.

LLMs Confirmed to Have a Recency Bias

The News: Two recent independent studies show that LLMs like ChatGPT are designed to give preference to newer information, even when the content is essentially the same.

In August, SEO consultant Metehan Dalgıç uncovered a setting in ChatGPT’s system called “freshness scoring,” which seemed to indicate that newer content received a boost in search results.

Now a peer-reviewed study from Waseda University has independently confirmed that leading AI models including GPT-4, GPT-3.5, LLaMA3, and Qwen exhibit a clear and measurable recency bias.

The study showed that simply changing the date on a piece of content, without changing anything else, was often enough to push it higher in the rankings. Some AI models were more biased than others, but the pattern was clear: old content gets buried.

What This Means: The implications are massive. Older content, even if it’s technically accurate and more authoritative, is being deprioritized by LLMs in favor of more recent content simply because it’s dated. To stay visible in LLM results, be sure to refresh your content regularly. If your strategy doesn’t prioritize keeping your most important information about your business up to date, it won’t just lag behind, it risks becoming invisible.

Google Provides Guidance on Review Extortion Scams

The News: “Review extortion scams” are schemes in which bad actors flood a business’s Google Business Profile with negative reviews, then demand payment, goods, or services in exchange for removing those reviews or stopping further attacks.

While review extortion scams have always been a problem for local businesses, they have grown more pervasive over the past few years, aided by a combination of increased AI automation, digital payment methods (often untraceable), and low legal enforcement (especially across borders).

To help combat this, Google has created a dedicated support page with guidance on how to handle review extortion. They advise businesses not to respond to or negotiate with the scammers, and instead to report the incident directly to Google using their new merchant extortion report form.

What This Means: Businesses are encouraged to include any evidence of the scam, such as screenshots and emails. Once submitted, Google will investigate and take appropriate action, although specific details may not always be shared due to privacy policies.