Discussion
  • Search Console
  • SEO Analytics

Query groups in Search Console Insights: useful for content planning, or too opaque to act on?

Published 1 reply

Discussion starter: an SEO measurement question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.

Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.

1 reply

  • admin
    Admin
    Original poster

    Useful for triage, too opaque to be the plan of record. Query groups are good at one thing the Performance report is bad at: showing topic-level movement across thousands of query rows at a glance. They are poor at the things content planning actually needs, such as a stable definition you can track, a view of demand you do not already capture, and a signal of whether two queries belong on one page or two.

    The way to get value from them is to treat each group as a hypothesis. When a group moves, rebuild it as your own filter in the Performance report, work out why it moved, and only then decide what to write, refresh or merge. Used that way, the opacity matters much less.

    What Google has said about how the groups work

    Google introduced Query groups as a card in Search Console Insights in October 2025. The points that matter for planning, as reported at launch:

    • The groups are computed with AI, and Google says they “may evolve and change over time”.
    • They are shown only to properties with a large volume of queries, on the reasoning that grouping matters less for sites with few queries.
    • The card reports each group’s clicks and its change against the previous period, so the main use is seeing which topics are rising or falling.
    • Launch coverage mentioned no API access and no query-group filter in the Performance report. Check the current documentation before building a reporting process around it.

    Google has not documented how membership is decided, whether anonymised queries contribute to a group’s totals, or how a period comparison handles queries that moved between groups. Those gaps are where the “too opaque” feeling comes from, and they are legitimate.

    Four limitations that matter for content planning

    Limitation Why it matters How to work around it
    Membership can change A group “trending up” might partly reflect queries joining the group, not more clicks on the same queries Log group names and visible member queries each month, and rebuild important groups as regex filters so you have a fixed definition
    Clicks carry the headline A fall in clicks can mean less demand, a lower click-through rate or lost rankings. Each needs a different response. Decompose in the Performance report: impressions, CTR and position for the rebuilt filter
    It only sees your existing footprint Groups are built from queries you already appear for. Topics you have no presence in never show up. Use keyword research and results-page review for gap discovery. Query groups cannot do that job.
    Similar meaning is not the same intent Two queries can mean similar things and still need different pages if Google ranks different page types for them Check the results pages, or the overlap in ranking URLs, before merging or splitting content

    Important distinction:

    A semantic grouping tells you which queries are about the same thing. A content decision needs to know which queries can be served by the same page. “Invoice template” and “how to write an invoice” are clearly related, but one usually rewards a downloadable template and the other a guide. Planning pages from semantic groups alone is how sites end up with one bloated page trying to satisfy two jobs, or two pages competing for one.

    A workflow that makes the groups actionable

    1. Record what you see. Once a month, note the largest, rising and falling groups and the member queries visible for the groups you care about. Because Google says groups may change, your own log is the only way to compare like with like later.
    2. Rebuild important groups as filters. Write a regex query filter in the Performance report that matches the group’s member queries. This becomes your stable definition. Search Console’s AI-powered report configuration can speed up the setup, but check the filter it produces against the actual queries.
    3. Decompose the movement. For the rebuilt filter, compare impressions, clicks, CTR and average position over the same periods. This step turns “group down 20%” into a diagnosis.
    4. Check known confounders. After Google stopped supporting the &num=100 parameter in September 2025, many sites saw desktop impressions fall and average position improve, while clicks were affected far less, according to Brodie Clark’s analysis. AI Mode data has been counted in Performance totals since June 2025. The March and May 2026 core updates also sit inside most current comparison windows.
    5. Map queries to pages. Switch to the Pages tab with the filter applied. If several URLs share the group’s clicks, or the ranking URL alternates, you may have cannibalisation rather than a demand problem.
    6. Decide, then annotate. Refresh, consolidate, create a new page, or do nothing. Mark the change with a Search Console custom annotation so the next comparison has context.

    Reading the patterns

    What the rebuilt filter shows Likely interpretation Verify with Likely action
    Clicks down, impressions stable, CTR down, position stable Results-page change, such as an AI Overview or new features above you Live results pages; the generative AI performance report for those pages Do not rewrite the content first. Review titles and snippets, and whether the page offers something an AI summary does not.
    Clicks and impressions down, position stable Falling demand or seasonality, or lost coverage of some queries Year-on-year comparison; which member queries disappeared Often nothing, if seasonal. If queries vanished, check what now ranks for them.
    Position down across the group Ranking loss, possibly competitive or update-related Dates against core updates; which pages overtook you Content and quality review of the ranking page
    Group rising, no page built for it Adjacent demand you are catching by accident Results pages for the top member queries A candidate for a dedicated page, if the intent differs from the page currently ranking
    Clicks split across several URLs Internal competition Pages tab under the filter, over time Consolidate or clarify the purpose of each page

    Hypothetical example:

    An accounting software site sees a group labelled around “invoice templates” trending up. Rebuilt as a regex filter, the growth sits almost entirely in queries for templates in specific file formats, served by one generic template page that has a low signup rate. Meanwhile a smaller group around “invoice payment terms” is falling, and the decomposition shows stable position with a falling CTR. The planning outcome is not “write more about invoices”. It is one set of format-specific template pages that route to the product, and a snippet and results-page review for the payment-terms guide before anyone rewrites it. (Invented scenario.)

    Where query groups fit in a content plan

    They are good for three things. First, spotting which existing topics are gaining or losing traction without scanning thousands of rows. Second, prioritising refreshes, because a falling group points you to a cluster of pages rather than one query. Third, explaining movement to stakeholders at topic level, which most people outside SEO find easier to follow than query lists.

    They are weak for three things. They cannot discover topics you have no footprint in, they do not size demand (clicks are not demand), and they do not tell you how to structure pages. For those jobs you still need keyword research, results-page review and your own clustering by ranking-URL overlap.

    One more practical point: separate branded from non-branded queries before drawing conclusions. A group dominated by brand-plus-topic queries tells you about brand demand, not about your content’s ability to win new searchers. The Performance report’s branded queries filter makes this check quicker than it used to be.

    The practical position

    I would check query groups monthly as an early-warning and idea list, then do the real analysis in the Performance report with filters I define and control. A group earns a place in the content plan only after its movement survives three checks: the rebuilt filter shows the same trend, the decomposition explains why, and the results pages confirm what kind of page the intent rewards. If a group’s movement disappears when rebuilt as a fixed filter, treat it as a change in the grouping, not in your search performance.

    Need help with this on your own site?

    Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.

    See SEO services

Add a reply

Replies are open to approved members. Log in or apply to join.