If organic traffic fell during a Google core update, don’t start editing pages yet. Most of the work is diagnosis: rule out causes that only look like an update, confirm the timing, find which page types lost visibility and see who took your place. The fixes that work are usually site-wide quality changes, and recovery often waits for a later update.
Ruling out tracking, technical, seasonal and AI Overview causes
Core updates get blamed for drops they didn’t cause. Before anything else, check the explanations that are cheaper to confirm and faster to fix.
| Cause | What it looks like | Where to check |
|---|---|---|
| Tracking break | GA4 organic sessions fall while Search Console clicks hold | Daily Search Console clicks vs GA4 organic sessions; tag deployments, consent banner changes |
| Technical issue | Sharp drop starting on a deploy date, often limited to certain templates | Page indexing report, Crawl stats, robots.txt, noindex and canonical tags, server errors |
| Seasonality | The same dip in the same weeks last year; lower demand overall | Year-over-year comparison in Search Console, Google Trends |
| AI Overviews | Impressions and positions roughly steady, CTR falling on informational queries | Search Console CTR by query; the live results page |
| Manual action or security issue | Sudden, severe, with a notice in Search Console | Manual actions and Security issues reports |
Start with Search Console clicks against GA4 organic sessions. Search Console counts clicks on Google’s side, so a drop that shows in GA4 but not in Search Console is a measurement problem, not a ranking problem. A removed tag, a tag that stopped firing on one template or a new consent banner can all produce it.
Then check branded queries separately. If branded clicks fell, demand for your name dropped, and no amount of SEO work on non-branded pages explains that.
If impressions held but clicks fell on question-style queries, the answer may now sit on the results page itself. That’s a different problem with a different fix, covered in losing clicks to Google AI Overviews.
Confirming the timing lines up with an update
Google lists core updates on the Google Search Status Dashboard with a start date and a completion date. Rollouts usually take around two weeks, sometimes longer, and rankings can swing during that window before settling.
Plot daily Search Console clicks and impressions for the 90 days around the rollout and look for three things:
- A drop that begins inside the rollout window, not days before it.
- Impressions falling along with clicks. That’s lost visibility, not just lost CTR.
- A step down that holds after the rollout completes, rather than a dip that bounces back.
Then check your own release log. If you shipped a redesign, a migration, a new template or a batch of pages the same week, you have two candidate causes and need to separate them before touching anything. Also check whether a spam update rolled out nearby; a drop tied to one calls for a different investigation.
Segmenting the loss by page type, intent and query
A site-wide average hides the pattern. Export Search Console data for two equal periods, for example the 28 days before the rollout started and the 28 days after it completed, and break the change down three ways.
By page type
Group URLs by template: blog posts, product pages, collection or category pages, location pages, comparison pages, docs. Search Console’s page filter handles simple URL patterns. On larger sites, export through the API or Looker Studio and group in a spreadsheet.
By intent
Tag the top losing queries as informational, commercial, transactional or navigational. A loss concentrated in informational queries while commercial pages held tells a very different story from a loss on product pages that drive revenue.
By query and position
Separate queries that slipped a few positions from those that fell off page one. Small slides across thousands of queries point to a broad reassessment of the site. A few head terms collapsing while the long tail holds points to specific competitors or a shift in how Google reads those queries.
A hypothetical example: a SaaS site loses a third of its organic clicks. Segmented, the product, pricing and integration pages are flat. Nearly all of the loss sits in a library of several hundred short glossary posts written to catch traffic. That isn’t a whole-site collapse. It’s one content type judged weaker than the alternatives, and it may be dragging on everything else.
Studying the pages that replaced you
Take the 20 to 30 queries where you lost the most clicks and look at what ranks now, using a rank tracker’s historical snapshots or a clean browser. For each, note:
- Who ranks: forums, publishers, brands, marketplaces, video
- The format: guide, list, tool, product page, comparison
- What they offer that you don’t: first-hand experience, original data, pricing, screenshots, a named expert, a faster answer
- Whether intent shifted: a query that used to return guides may now return product pages, or the reverse
Sometimes the answer is uncomfortable: the new page is simply better. Sometimes it’s a shift you can’t match, such as Google now favoring forum discussions for a query. Know which before deciding what to fix. And don’t copy the winner’s word count or headings. Work out what problem it solves better than you do.
What site-wide quality problems look like in practice
Google says its ranking systems work mainly at the page level but also use some site-wide signals. In practice, a large share of weak pages can hold back strong ones. When I audit a site after a drop, these are the patterns I look for:
- Thin or near-duplicate pages at volume: tag archives, programmatic location pages that swap city names, product variants with identical copy
- Content written for keywords, not readers: topics unrelated to what the business does, published to catch search traffic
- Summaries of what already ranks: no examples, no data, no experience, nothing a reader couldn’t get from the current top result
- Unclear ownership: no author, no about page, no sign of who stands behind the advice
- Stale pages presenting old prices, features or advice as current
- Answers buried under long intros, or crowded out by ads and popups
One or two of these on a small corner of the site rarely matters. Hundreds of URLs in a pattern usually does.
Pruning, consolidating and improving
Build an inventory of every URL in the affected page types, with 12 months of clicks, impressions, backlinks and conversions. Then decide page by page.
| Action | When | How |
|---|---|---|
| Improve | The topic matters to your business, the page still earns impressions and can be made clearly better | Fix intent, add first-hand detail, examples and current facts |
| Consolidate | Several pages cover the same topic and split the signals | Merge into the strongest URL and 301 redirect the others |
| Noindex | Useful to visitors on the site but has no search value | Keep it live with a noindex tag |
| Remove | No traffic, no links, no business reason to exist | Return a 404 or 410; redirect instead if it has links |
Remove and consolidate first. It shrinks the rewrite list and focuses effort on pages that deserve it. When the problem sits in a template, such as product pages running manufacturer copy, fix the template instead of the pages one by one. The rewrite process for pages worth keeping is covered in content refresh strategy.
This diagnose-then-fix sequence is how I run technical SEO and content systems work after a drop: one inventory, one decision per URL, shipped in batches by page type so the effect can be measured.
Why recovery often waits for a later update
Google’s guidance is that improvements can show up between core updates, because ranking systems update continuously. But larger reassessments tend to arrive with the next core update, and those come a few times a year. A realistic plan assumes months, not weeks.
Recovery isn’t guaranteed either. If competitors improved or Google now reads the query differently, the old rankings may not return even after good work. In that case the goal shifts to winning the queries where you can be the best result.
Keep leadership informed with leading indicators while you wait:
- Count of thin or duplicate URLs removed, merged or noindexed
- Improved pages recrawled, confirmed with URL Inspection or crawl logs
- Impressions and average position on improved pages vs untouched pages
- Non-branded clicks by page type, reviewed monthly
- Organic conversions on commercial pages, so the business impact stays visible
What not to do after a drop
- Don’t change anything during the rollout. Rankings fluctuate until it completes, and you’ll react to noise.
- Don’t disavow links or buy new ones as a fix. Core updates aren’t about link spam.
- Don’t change publish dates without changing the substance of the page.
- Don’t mass-produce replacement content with AI. Swapping thin pages for more thin pages repeats the problem at speed.
- Don’t launch a redesign or change URLs as a recovery move. It adds new variables and new risk.
- Don’t tweak titles and keyword placement page by page, hoping to reverse a site-level reassessment.
- Don’t delete pages that earn links or conversions just because they lost rankings.
- Don’t judge the result after two weeks. Measure at the next core update.
Get it built
If traffic fell and you can’t tell whether it’s the update, your tracking or your site, the Growth Audit separates the causes and ends with a prioritized fix list. It’s $1,500 fixed and credited if we continue. See pricing or get in touch.