Discussion
  • Agency

A client drops after a core update despite solid work. What do you report, and how do you avoid blaming the algorithm by default?

Published 1 reply

Discussion starter: an agency 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

    Report the drop as a finding with a confidence level, not as a verdict. The client needs to know what fell, where, from when, what else happened in the same window, what you have ruled out, what the pattern points to, what it costs the business, and what you will do next, by when. “We were hit by the core update” answers only one of those, and it is usually the one you are least sure about.

    The key distinction is that a core update is a timestamp, not a diagnosis. Even when the timing is clean, the useful question is not “did the update cause this?” but “what did Google decide it now prefers for these queries, and why aren’t we that?” Solid work can still lose, because core updates judge relevance and usefulness against competing pages. Whether your team completed its checklist is not part of that assessment.

    Why “solid work” and “lost the core update” can both be true

    An analogy helps. Picture an exam re-marked against a revised rubric, with every student’s paper re-ranked at the same time. Your essay has not changed, and it has no spelling mistakes. But the marker now reads the question slightly differently and weights some qualities more heavily, and a few other students rewrote their answers. Your rank moves even though nothing on your page did.

    Technically, Google describes core updates as broad reassessments of content. Its documentation makes the point that pages that move down are not necessarily bad; other pages were judged a better fit. Aleyda Solis’s analysis of the May 2026 core update found the biggest visibility shifts were driven by intent, market fit and source type. Forums and Q&A sites lost UK visibility, transactional marketplaces gained, and UK results moved from US retailers towards local-market domains. If Google decides a query is now better served by a marketplace than a brand’s buying guide, that is not a judgement on the guide’s craft. It is still your problem to respond to.

    So “solid work” needs unpacking in the report. Technical hygiene, publishing cadence and link acquisition are inputs. A core update assesses the output: is this page the most useful answer for this query, as Google now reads it? A team can do the inputs well and still produce pages that are interchangeable with competitors. That is uncomfortable to write down, but it is often the most useful line in the report.

    Timing alone cannot prove the cause

    Update windows overlap with each other and with your own releases. Here is the 2026 calendar from the Google Search Status Dashboard:

    Update Start End
    February 2026 Discover update 5 Feb 27 Feb
    March 2026 spam update 24 Mar 25 Mar
    March 2026 core update 27 Mar 8 Apr
    May 2026 core update 21 May 2 Jun
    June 2026 spam update 24 Jun 26 Jun
    August 2026 spam update 18 Aug 21 Aug
    September 2026 spam update 24 Sep Still listed as in progress at the time of writing

    A drop that started on 25 March 2026 cannot be the March core update, which began two days later. Google also documents that it runs smaller core updates on top of the announced ones: “We’re continually making updates to our search algorithms, including smaller core updates”. So a movement with no announcement behind it is not proof that the cause is on your site either. Timing narrows the candidates. It does not choose between them.

    What I would check before writing the report

    1. Measurement first. Did tagging, consent settings, GA4 configuration or the Search Console property change? Do Search Console clicks and analytics sessions fall together? The &num=100 change in September 2025 showed how impressions and average position can shift sharply, while clicks were affected far less, according to Brodie Clark’s analysis.
    2. Exact timeline. Find the first day of decline at daily granularity. Overlay update windows, deployments, CMS or template changes, content changes, migrations and seasonality. Google suggests waiting at least a full week after a core update completes, then comparing that week with the week before the rollout started.
    3. Segmentation. Split branded and non-branded queries with the branded queries filter. Then split by directory or template, intent class, country, device and search appearance. Check Discover separately.
    4. Non-algorithm Google causes. Google’s own debugging guide lists technical issues, security issues, spam policy violations and manual actions, seasonality and site moves alongside algorithm updates. Check the Manual actions, Security issues, Page indexing and Crawl stats reports.
    5. Who replaced you. For the top lost queries, record which URLs now rank and what type of source they are: forum, marketplace, manufacturer, publisher, video or local domain. Note SERP layout changes such as AI Overviews.
    6. Business impact. Map lost clicks to organic conversions, leads and revenue by page group. A large traffic loss on low-intent informational pages and a small loss on product pages are very different conversations, and the smaller one may cost more.

    Reading the pattern

    Pattern What it suggests What to verify How to word it
    Decline starts before the update window, or on a release date A site change is the more likely trigger Release log, rendering, internal links, canonicals, crawl stats “Timing points to our release on [date], not the update.”
    Branded flat, non-branded down across many templates Broad reassessment consistent with a core update Peer domains, source types now ranking “Consistent with the core update. Confidence about the trigger: high. About the reason: moderate.”
    Losses concentrated in one directory or template A page-type problem such as thin programmatic pages, stale content or an intent mismatch What ranks now for that template’s queries “The update exposed a weakness in [page type].”
    Positions stable, clicks down SERP layout or CTR change, not a ranking loss SERP samples, the Search Console generative AI impressions report “Rankings held. Fewer searchers are clicking.”
    Replaced by a different source type Intent or source-type shift. On-page tweaks are unlikely to reverse it. Whether your format can serve the new intent “Google now favours [forums or marketplaces] for these queries. This is a strategy decision, not a fix.”
    Drop sits inside a spam update window Spam policy exposure, possibly through links or hosted third-party content Manual actions report, link acquisition history, third-party sections “We are checking spam policy exposure before anything else.”

    What the report should contain

    1. Summary with confidence labels. Separate what is verified, what is likely, what is a hypothesis, and what is still unknown.
    2. Impact in business terms. Lead with leads, revenue or pipeline, then traffic.
    3. Timeline. Show the drop against update windows, releases and seasonality on one chart.
    4. Where it happened. The segmentation findings: which templates, queries, countries and devices.
    5. What was ruled out, with evidence. This section is what separates analysis from blame.
    6. Ranked hypotheses. Include evidence for and against each, including anything your own work may have contributed.
    7. Actions, owners and dates. Also list what you are deliberately not doing yet, such as mass rewrites or deleting pages.
    8. A checkpoint. Give the date you will reassess and the result that would change the plan.

    The wording matters. Compare “The May core update hit us” with this template: “Non-branded clicks to [section] fell [x]% in the week after the May 2026 core update completed, compared with the week before it began. Branded clicks were flat. The decline is concentrated in [n] templates. No deployments, tracking changes or manual actions coincide. For [y] of the top [z] lost queries, the new top results are [source type].” The second version gives the client something to check, and gives you something to be wrong about in a useful way.

    Hypothetical example:

    A client’s non-branded clicks start falling on 25 March 2026, and the first draft of the report says “March core update”. The dashboard says otherwise: the core update began on 27 March, and the spam update ran from 24 to 25 March. The release log also shows a blog template change on 24 March that moved related-article links behind a JavaScript tab. The honest report lists three candidates in order. It checks rendered HTML and crawl stats for the template change, reviews spam policy exposure, and only then looks at movement during the core update window. It might still end at “core update”, but now the client can see why.

    Habits that stop “the algorithm” becoming the default answer

    • Keep a change log and annotate it. Search Console now supports custom annotations, so releases and content changes sit on the same chart as the traffic.
    • Complete the non-algorithm checklist first. Do it before the word “update” appears in any draft.
    • Watch a peer set. Track 5 to 10 comparable domains. If the whole sector moved, the cause is probably systemic. If only your client moved, look inward first.
    • Separate cause from response. Even when the update is the trigger, the response is still yours to own.
    • Do not overcorrect. Inventing a site-side cause so you look proactive is the mirror image of blaming Google. Rushed sitewide rewrites or pruning straight after a drop also destroy your ability to measure what helped.

    What happens next, and how to test it

    Set expectations early. Google says recovery could take several months for its systems to recognise sitewide improvements, and states that “there’s no guarantee that changes you make to your website will result in noticeable impact in search results”. Improvements can be picked up between announced updates, so you are not simply waiting for the next one.

    Test instead of rewriting everything. Pick the page group with the biggest revenue loss. Write a specific hypothesis, for example: “these guides lost to marketplace listings because they lack current pricing and first-hand comparison”. Improve a subset and hold back a comparable group. Measure non-branded clicks and conversions for both over at least one quarter, noting any update windows that fall inside it. If the improved group recovers and the holdout does not, scale the change. If both recover, the update may have partly reversed, and you learned the change was not the lever. If neither moves while the new source type keeps winning, the conversation with the client is about whether to compete in that format at all.

    The decision rule I would use: name the core update as the trigger only after measurement, releases, seasonality, spam and technical causes are ruled out in writing. Even then, name what Google now prefers for the lost queries and say plainly whether the client’s pages can become that. Clients rarely lose trust over a drop. They lose it over an explanation that turns out to be wrong.

    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.