Diagnosing an organic traffic drop in an hour

Organic traffic falls, someone notices, and within an hour there are five theories: an algorithm update, a competitor, the new design, a penalty, AI answers. Each theory leads to a different set of changes, and if the team acts on all of them at once, nobody will ever know what worked. A fixed diagnostic order avoids that. Most drops have ordinary causes that can be confirmed or ruled out quickly, and the sequence below goes from the cheapest checks to the most complex.

If the trigger was a falling score in a technical monitor rather than falling traffic, start with the shorter checklist in your score dropped: what to do.

Minutes 0–10: is the drop real?

1. Compare Search Console with analytics. If Search Console clicks are steady but analytics sessions fell, search is fine and measurement changed: a tag removed in a redesign, a consent banner update, a filter in analytics, a change in channel definitions.
2. Compare year over year. A December drop for a garden centre is the calendar. Use the same weeks last year, aligned by weekday.
3. Check the data delay. The last two days in Search Console are often incomplete and revised later. End your comparison window three days ago; see the Performance report guide.
4. Size it. A 5% drop over a week is often noise. A 30% drop starting on a specific day is a signal. Note the exact start date; almost every later step depends on it.

Minutes 10–25: locate the drop

A drop is almost never spread evenly. Find where it lives, using equal before and after periods:

Cut	Where in Search Console	What it suggests
By page or section	Pages tab; page filter by folder	One template or section: technical or content change there
By query type	Queries tab; brand vs non-brand regex	Brand drop: demand or reputation. Non-brand: rankings or SERP changes
By device	Devices tab	Mobile-only: mobile template, speed, rendering
By country	Countries tab	One market: local competition, hreflang, regional issue
By search appearance	Search appearance tab	Lost rich results: structured data problem
By search type	Search type selector	Image or video traffic dropped separately from web

The regex filters guide has patterns for brand and section filters. Write down where the loss is concentrated before going further. "Product pages on mobile in the UK" is a diagnosis you can work with; "traffic is down" is not.

Minutes 25–40: rule out technical causes

Technical faults are the most common fixable cause and the most urgent, so check them next:

• Indexing: in the Page indexing report, look for a jump in "Excluded by noindex", "Blocked by robots.txt", "Server error", "Not found" or "Duplicate, Google chose different canonical" around the date of the drop.
• robots.txt and noindex: check the live robots.txt and the meta robots of a few affected pages. A release that added a sitewide noindex is a classic.
• Canonicals and redirects: did a template start pointing canonicals elsewhere, or did URLs change without redirects?
• Speed and availability: server errors or timeouts in logs; Crawl Stats in Search Console for a rise in errors or response time.
• Rendering: for JavaScript sites, use URL Inspection's live test on an affected page to confirm the content is present; see the URL Inspection guide.
• Recent changes: list every deploy, plugin update, CMS change, migration or CDN change in the two weeks before the drop.

A scheduled technical check makes this step fast. If your monitoring shows the score fell on the same day as traffic, the cause is almost certainly technical. The free SEO checker covers the home page basics in under a minute.

Minutes 40–50: look at the results page

If positions held but clicks fell, the results page probably changed:

• An AI-generated summary now appears above the results for your main queries; see AI Overviews.
• New ads, shopping units, video carousels or local packs push organic results down.
• You lost a featured snippet or rich result you previously had.
• A competitor's result became more attractive (better title, review stars, prices).

Search your top lost queries and compare with what you remember or with screenshots from rank-tracking tools. A falling CTR at a stable average position is the numeric signature of this cause.

Minutes 50–60: updates, competitors and demand

• Algorithm updates: does the drop start within an announced core or spam update rollout, and affect many pages across the site? If so, follow the process in the core updates guide rather than changing things mid-rollout.
• Competitors: for the top lost queries, who now ranks above you? A new strong page from a competitor explains a drop for specific queries, not for a whole site.
• Demand: if impressions fell for a whole topic, including queries where your position did not change, fewer people are searching. Search trend tools confirm it.
• Manual actions and security: check both reports; they are rare but decisive. See manual actions.

Gradual declines are a different problem

Everything above assumes a drop with a start date. Many sites face something harder to see: a slow slide of a few percent a month over a year. Gradual declines rarely have a single cause. They usually come from content ageing while competitors improve, from shifts in what the results page shows, and from internal structure eroding as the site grows. The diagnosis is different: compare the same months year over year by page, list the pages with the largest absolute losses, and treat it as a content and structure review rather than an incident. The content decay guide covers that process.

Communicating while you investigate

Traffic drops make people anxious, and anxious people make changes. A short message early in the hour helps: "Organic clicks fell about a quarter from Tuesday. We're checking tracking, technical changes and the results page now; next update by end of day. Please hold off on site changes until then." It costs nothing, and it prevents the well-meaning title rewrite or plugin update that would muddy the evidence.

Write it down

End the hour with a short note:

1. What dropped (metric, size, start date).
2. Where (pages, queries, device, country).
3. Ruled out (tracking, seasonality, technical items checked).
4. Most likely cause and the evidence for it.
5. Action, owner and the date on which you will check the numbers again.

This note prevents the most expensive mistake in drop diagnosis: making several unrelated changes at once. It also gives the next person a head start if the drop returns.

A worked example

A hypothetical home-goods shop sees organic clicks fall about a quarter week over week. Search Console confirms the drop, and last year shows no such dip in the same weeks. By page, almost all the loss is on product pages; categories and guides are flat. By device, it is mobile only. The Page indexing report shows no change, but URL Inspection's live test on a product page renders an empty product area on mobile. A deploy four days before the drop introduced a mobile-only lazy-loading script for product details that fails for crawlers. The fix is rolled back the same day. Clicks recover over the following two weeks as pages are recrawled. Without the cut by page and device, the team would have been reading about core updates.

Common traps

• Blaming an update because one happened to be announced that week.
• Reading the last two days of Search Console data as a fall.
• Looking only at totals, never at segments.
• Changing titles, content and templates in the same week as the investigation.
• Forgetting that analytics and Search Console measure different things.
• Stopping at the first plausible cause. A deploy and an update can coincide; check that the evidence fits the pattern of the loss, not just the date.