Google rankings dropped? Work through this in order
A decision tree that separates the four causes of a traffic drop (reporting, technical, content and environment) before you change anything.
Entrovix AIAug 18, 2026 12 min read
Traffic falls and the instinct is to start fixing. That instinct is the problem: a drop has four possible causes with four unrelated remedies, and acting before you know which one you have usually means changing a page that was never at fault.
The sequence below narrows it down in about twenty minutes, and most of that is reading reports you already have.
Start with the shape of the drop
Open Search Console → Performance and set the range to the last 6 months. Before anything else, establish three facts: when it started, whether it was a cliff or a slope, and whether it affected everything or one section.
| Shape | Usually means |
|---|---|
| Sudden cliff on one date | Something you deployed, or a manual action |
| Gradual slope over weeks | An algorithm update, or competitors improving |
| One section only | A technical fault scoped to that template |
| Everything at once | Site-wide: indexing, hosting, or a domain-level change |
| Weekend or seasonal dip | Often not a drop at all |
The decision tree
Four questions in this order. Each one eliminates a whole category of cause.
1. Did impressions fall, or only clicks?
This split is the most informative single thing in the report, and it is the step most often skipped.
- Impressions steady, clicks down: you are still ranking. Something changed about the result itself: a rewritten title, a competitor's better description, or a SERP feature taking the attention above you.
- Impressions down, average position steady: fewer people are searching, or the query mix changed. That is demand, not a fault.
- Impressions down and position down: you genuinely lost rankings. Continue to question 2.
- Everything at zero from one date: this is a tracking or indexing failure, not a ranking one. Check that analytics is still installed and that the pages still return 200.
2. Are the pages still indexed?
Run URL Inspection on two or three pages that lost the most. If they are no longer indexed, stop. This is not a ranking problem and no content work will help. It is a technical fault, and how to check whether Google indexed your site covers the statuses and what each one means.
The causes that produce a sudden site-wide de-indexing are short and specific: a noindex that shipped with a deploy, a robots.txt replaced by a staging version, a canonical template pointing everywhere at one URL, an expired certificate, or a server returning 5xx while Google was crawling.
3. Did anything change on your side?
Compare the drop date against your own history, which is the step people skip because they assume they would remember. Check deploy logs, CMS revision history, plugin updates and DNS changes.
- A redesign or migration. The commonest cause of a large drop, and the cause is nearly always an incomplete redirect map or content trimmed for visual reasons.
- URL changes without redirects. Every changed URL that was earning traffic loses it entirely.
- A plugin or theme update that altered titles, canonicals or robots directives sitewide.
- Hosting or CDN changes producing slow responses or intermittent errors during crawls.
- Content removed because it looked untidy. Thin-looking pages sometimes earn a great deal.
4. Did the environment change?
Only reach this question after the first three come back clean. Check your drop date against the confirmed Google core update dates (opens in a new tab) — 2026 has had two, in March and May.
If the dates line up, the cause is a systemic reassessment of your site rather than a fault on one page, and there is no switch to flip. Google's own guidance on core updates is that there is nothing to fix in the technical sense; the response is to make the content genuinely more useful and wait for the next assessment.
If the dates do not line up, stop blaming the algorithm. A drop on a day with no confirmed update is far more likely to be something on your own site, and question 3 is where to look again more carefully.
Causes people reach for that are usually wrong
| Assumed cause | Why it is usually not that |
|---|---|
| A competitor attacked us with bad links | Google largely ignores obvious spam links. Genuine negative SEO is rare |
| We were penalised | Manual actions appear in Search Console. If there is nothing there, there is no penalty |
| The site is too slow now | Speed is a weak tiebreaker. It does not cause a cliff |
| We need more content | Publishing more rarely recovers a drop, and can compound it if the pages are thin |
| Someone copied our content | Duplicate content is not a penalty. Google picks a canonical, usually the original |
What to do once you know the cause
| Cause | Response | Realistic recovery |
|---|---|---|
| Technical or indexing fault | Fix and request re-crawl | Days to weeks once recrawled |
| Migration with broken redirects | Complete the redirect map from old URLs to closest equivalents | Weeks |
| Core update | Improve the content genuinely; wait for the next update | Months, and not guaranteed |
| Lost to better competitors | Read what now outranks you and close the gap | Months |
| Falling search demand | Nothing. It is not a fault | Not applicable |
Before changing a single page, be able to say which of the four causes you have. The most expensive mistake in this whole area is rewriting content to fix a redirect.
Questions people actually ask
How much of a drop is normal noise?
Day-to-day movement of 10 to 20 percent is ordinary, and single keyword positions fluctuate constantly. Compare 28-day periods rather than days, and only investigate a sustained change in the trend.
How long before I should worry?
If a drop persists beyond two weeks and the shape is a cliff, investigate immediately, because that pattern is nearly always technical. A slow slope over a quarter is worth a content review rather than an emergency.
Should I disavow links?
Almost certainly not. The disavow tool exists for sites with a manual action for unnatural links. Using it speculatively can remove links that were helping, and Google has been consistent that most sites never need it.
Can I recover from a core update?
Yes, but not quickly and not by tweaking. Recovery generally follows a later update, and only for sites that meaningfully improved in between. Treat it as a quarter of work on content quality, not a week of adjustments.
