Did your Google Search Console Impressions Just Drop? A Recent Google Update May Be the Answer
Google has disabled the &num=100 results parameter, which has been used to display up to 100 search results on a single page. This change is affecting many SEO tools and could explain why many SEO professionals have seen a drop in impressions starting September 10th, with average positions increasing accordingly.
While Google has not provided an official reason, a spokesperson has stated that the &num=100 URL parameter is “not something that we formally support.”
Over the past year, SEO professionals have been grappling with rising impressions that don’t translate into clicks – also known as the “great decoupling” – often attributing it to AI Overviews. This recent change to the &num=100 parameter, however, suggests that tool-related bot activity might have been a hidden cause all along, especially on desktop.
To further complicate tracking in Search Console, the August 2025 spam update finished rolling out on September 22nd and took 27 days to complete. Volatility in the SERPs was heated around September 9th, closely coinciding with the observed changes to impressions on the 10th. If you think your website was affected by the spam update, reach out to your ROI Revolution SEO team for assistance.
AI Mode Will Be Added to the Chrome Address Bar
Chrome’s address bar, also called Omnibox, will be getting its own AI Mode search button later this month, according to a recent Google announcement. This means that as the user types a search query into the address bar, they will have the option to opt for an AI Mode experience over a traditional Google search. This evolution may support some of our SEO team members’ theory that the search giant is testing the waters with AI Mode before rolling it out as the primary search experience, but this is not confirmed.
Here’s what the updated Omnibox looks like:
Google Search Quality Rater Guidelines Updated
The Search Quality Rater Guidelines are the criteria provided to Google’s human quality raters as they evaluate the quality of webpages and search results. This documentation is updated regularly, but two key changes stood out to SEOs:
Updated the YMYL Society definition – Google expanded its classification for Your Money Your Life (YMYL) topics related to ‘society.’ The added language is identified below in bold text, and removed language is identified with strikethrough.
“YMYL Government, Civics & Society: Topics that could negatively impact groups of people, issues of public interest, trust in public institutions, election and voting information, and any other informational topics about government, civics or society that impacts people’s lives. etc.”
Added AI Overviews examples – Multiple examples of AI Overviews results (along with corresponding queries, user intent, and rating guidelines) were added throughout the documentation. These examples will likely prove helpful for SEOs and brands as we figure out how to optimize to show up more frequently in relevant AI Overviews.
The full document is enormous, but the most recent version of the SQR Guidelines is publicly accessible if you’re curious (and have plenty of spare time on your hands).
Managing Your Google Reviews is More Important Than Ever
When a user searches for a local business, the Google AI Overview search result may now directly pull and highlight individual reviews from the company’s Google Business Profile. This new development makes positive reviews more critical than ever, as they could be the first thing a user sees at the top of the search results.
AI Overviews is not the only place where reviews are a key feature of the search results. AI models are increasingly using signals like third-party mentions and reviews to evaluate a brand’s trust and authority.
DOJ vs Google Results, Plus More Legal Trouble?
Early this month, U.S. District Judge Amit Mehta determined that the fairest way to remedy Google’s monopoly is to force Google to share parts of its search data with rivals. The ruling was lighter than expected after months of speculation that Google’s parent company, Alphabet, might be forced to sell Chrome or end contracts that make Google the default search engine in many browsers and devices.
Google celebrated the win, while still vowing to appeal the original decision from 13 months ago that ruled that the tech giant is a monopoly. Investors happily sent the stock up 3%, and the Department of Justice stated it was a win for the American people.
During arguments over the last several months, more court documents revealed additional information about how Google’s search engine works:
- Google’s FastSearch is faster than Google Search and primarily used to ground and validate Gemini’s responses to user queries.
- Google incorporates user data in every stage of the search process, “from crawling and indexing to retrieval and ranking.”
- “Glue” is a log that collects data and user interaction about a query and its response.
- The deep-learning model RankEmbeded BERT uses 70 days of search logs and human quality raters scores.
- Google utilizes its spam score, page quality, and popularity signals to determine what to crawl and how often.
- While arguing against being forced to divest the Chrome browser, Google’s lawyers argued that “the open web is in rapid decline,” despite the company’s CEO, Sundar Pichai, stating in May that the web is thriving.
A new lawsuit filed by Penske, Rolling Stone’s publisher, against Google pointed to that last point as a reason for the filing. The lawsuit is the first one by a major American publisher over Google’s AI Overviews, and the complaint is that the SERP feature siphons traffic from their site, causing a decline in advertising and subscription revenue.
In response to the lawsuit, Google stated, “With AI Overviews, people find Search more helpful and use it more, creating new opportunities for content to be discovered. We will defend against these meritless claims.”
To add insult to injury (the light slap on the wrist from Judge Mehta), The European Union fined Google $3.5 billion dollars for being anti-competitive and favoring its own technology over others. Google will also be required to divest part of its ad-tech business. The company has 60 days from the September 4th ruling to inform the commission how it plans to comply.
Google Says Users Prefer Summaries
During an AI summit in New York City, Markham Erickson, Google’s vice-president of government affairs and public policy, said that users are shifting toward a preference to contextual summaries over “factual answers.” The statement was made in response to a question about the Penske lawsuit mentioned above.
Erickson said that Google’s goal is to maintain a “healthy ecosystem” with AI summaries and the original blue links of search results.
Like it or not, the statement reflects the same sentiment we found in our own survey this summer. Despite 61% of respondents occasionally or often seeing inaccuracies in AI-generated answers, 60% said that AI Overviews made Google Search better.
LLMs Do Not Directly Use Structured Data
In the ongoing debate about the value of schema markup, or structured data, some people are testing whether the AI search bots, like ChatGPT and Gemini, directly use structured data. The answer appears to be no. In the process of reading the page, the bots tokenize and destroy the markup.
However, in the effort to validate their answers with fresh information while streamlining resources, the AI search bots are using Google and Bing to search the web, while reserving resources for reasoning and logic. Google and Bing use structured data, so we’d like to remind you that indirectly, AI search is still using structured data (and you should, too).
SEO Still Isn’t Dead
In August, data from SimilarWeb showed that 95.3% of ChatGPT users also used Google. By comparison, 14.3% of Google users visited ChatGPT. While those proportions will likely change in the future, Google is still the dominant gateway to information. (And if the chatbots are using it, we’re not counting it out any time soon.)
ChatGPT Improves Its Search Functionality
Speaking of the popular tool, ChatGPT’s release notes on September 15 touted improvements to factuality, meaning fewer hallucinations when answering queries, formatting for quick understanding, and shopping. Now the tool is better at detecting shopping intent and showing products when users want them.
ROI Answers: SEO-Friendly Internal Site Search Tools
Recently, an ROI client asked us for our thoughts on choosing an SEO-friendly internal site search tool. While there are plenty of paid tools and plugins to choose from, here are our top factors to consider when selecting one for your site:
- URL Indexing Control: The way the internal search tool handles URLs is critical for SEO. The tool should give you control over the results page indexing – we recommend adding noindex tags for user-generated search results pages to prevent the creation of many low-quality, thin content pages that could compete with your high-value URLs for relevant rankings.
- Note that /search URLs should not be excluded in the robots.txt file if they are tagged with noindex/nofollow instructions, or you may inadvertently cause confusion for crawlers.
- Additionally, look for tools that generate clean, static, and descriptive URLs for results pages (ex. www.yoursite.com/search/navy-dress-shoes) rather than one that creates messy, parameter-filled URLs.
- Analytics Capabilities: The search queries your customers use are a goldmine of data for optimizing your site. Internal search tools should provide detailed analytics and reporting on what users are searching for, including the most popular queries, searches that return no results, and queries that lead to a purchase. This data can be used for many things, including the following:
- Identify content gaps and opportunities for new pages.
- Understand which areas of the site users may have trouble locating, which may lead to navigation and internal linking adjustments for better UX.
- Refine on-page SEO by incorporating the language your customers use.
- Inform paid search strategy by discovering new high-value keywords.
- Relevance & Weighting Settings: The primary goal of any search tool is to deliver accurate and highly relevant results. Some internal site search tools use algorithms to understand user intent, even with misspelled words, synonyms, or complex long-tail queries. Another advantage of advanced search tools is the ability to weigh pages differently, so specified URLs show up first in the results, helping you drive traffic to high-priority site areas.
- Robust Filtering Capabilities: E-commerce sites often have large product catalogs, so robust filtering and faceted search capabilities are an important consideration. Some site search tools can help users narrow down their search results by attributes like price, size, color, brand, or category.
- Speed & Performance: Site speed is important not only for user experience, but it’s also a ranking factor for Google and other search engines. Any internal search tool should load quickly. Features like “search as you type” can provide near real-time results as the user types in their query, and “autocomplete” capabilities can improve UX by speeding up the actual query input process.
ICYMI: Recent SEO Blog Posts
Get Featured in AI Overviews Search Results
Future-Proof Your Search Visibility: Introducing AI Overview Optimization
