Why did my Google rankings drop? A 30-minute checklist
Traffic falls and the first instinct is to blame an update. Nine times out of ten it is something narrower and fixable. Here is the order to check things in, so you find the cause before you start rewriting anything.
A drop feels like an emergency, which is exactly why most people fix the wrong thing. They rewrite a page that was never the problem, or they add 800 words to something that lost traffic because a competitor got a featured snippet.
Work through this in order. Each step either finds the cause or rules out a whole category, and the whole thing takes about half an hour.
First: is it actually a ranking drop?
Three different things look identical in an analytics chart:
- Rankings fell. Your average position got worse. Impressions usually fall with it.
- Demand fell. You rank exactly where you did; fewer people are searching. Seasonal topics and news-driven topics do this constantly.
- Clicks fell but positions did not. Something changed on the results page — an AI Overview, a new ad block, a video carousel — and it is eating the clicks that used to be yours.
Open Search Console → Performance → Search results, set the date range to Compare → last 3 months vs previous period, and turn on all four metrics. Now look at impressions and position together:
| Impressions | Position | What it means |
|---|---|---|
| Down | Worse | A real ranking drop. Continue below. |
| Down | Same | Demand fell, or you lost pages from the index. Check Pages → Indexing first. |
| Same | Same | A SERP feature is taking your clicks. Skip to step 3. |
| Up | Worse | You started ranking for more, weaker queries. Often not a problem at all. |
Do not trust a third-party rank tracker for this. Those tools check from one location, on one device, usually not logged in. Search Console is the average across everyone who actually saw you. When the two disagree, Search Console is right.
Step 1 — find the scope in Search Console
The single most useful question is how much of the site is affected, because the answer points at completely different causes.
- In the same Compare view, click the Pages tab and sort by the click difference.
- Count how many pages lost meaningfully — say more than 20% of their clicks.
Then read the shape:
- One page. Something about that page or that result changed. Steps 3 and 4 will find it. This is the most common case and the easiest to fix.
- A group of pages on one topic. Usually a competitor published something much better, or you have two of your own pages fighting for the same query (cannibalisation).
- Almost everything, on one day. Technical, or a core update. Step 2, then step 5.
- Almost everything, slowly over weeks. Usually content decay — nothing broke, the rest of the web simply got fresher than you.
Write the date the drop started somewhere. You will need it in step 5.
Step 2 — rule out the boring technical causes
These are unglamorous and they cause more drops than any algorithm update. Ten minutes, in this order:
- Is the page still indexed? Search Console → URL inspection → paste the URL. If it says "URL is not on Google", nothing else matters until that is fixed.
- Did a
noindexappear? SEO plugins, staging-to-live copies and theme updates all do this by accident. View source and search fornoindex. - Did robots.txt change? A single
Disallow: /pushed live on a Friday is a classic. Compare it with what you remember, and check the site is not still carrying rules from a redesign. - Did the URL change? Slug edits, category changes, a migration to a new permalink structure — if the old URL does not 301 to the new one, the rankings do not move either.
- Is the canonical pointing somewhere else? Self-referencing is normal. Pointing at a different page hands your rankings to that page.
- Is the site actually up for Googlebot? Check Search Console → Settings → Crawl stats for a spike in 5xx responses or a collapse in crawl requests. Aggressive firewalls and "security" plugins block Googlebot more often than anyone admits.
- Did the page get much slower? A plugin, a new ad script or a bloated hero image. Run PageSpeed Insights on mobile.
If any of these is the cause, fix it, then request indexing on the URL. Recovery is usually days, not months.
Step 3 — look at the results page itself
Open an incognito window and search the query yourself. You are looking for what changed around your result:
- An AI Overview at the top. It pushes the first organic result far down the page. Positions are unchanged; clicks are not. The counter is to be the source it cites and to answer questions it does not cover.
- A featured snippet someone else now owns. Look at the format they used — a short definition paragraph, a numbered list, a table. That format is the ticket, not the word count.
- More ads than before. Commercial queries grow ad blocks over time. Nothing you can do except be worth scrolling to.
- A People Also Ask block right under position one. Those questions are also a content map: if your page answers all of them, it deserves the snippet.
- Intent changed. If the results used to be blog posts and are now product pages or tools, Google has decided the query means something else. No amount of rewriting a blog post fixes that — you need a different page type, or a different query.
This step alone explains most "my position is the same but my traffic halved" cases.
Step 4 — read the page that took your spot
Not skim. Read it, with your own page open beside it, and answer four questions honestly:
- What does it cover that you do not? List the subheadings it has and you lack. This is usually the whole answer.
- How current is it? Check the visible date and, more importantly, whether the content is current — version numbers, prices, screenshots, this year's changes.
- Does it answer the query faster? Count the words before the actual answer. If theirs answers in the first two sentences and yours takes four paragraphs of preamble, that is a real ranking factor dressed up as writing advice.
- What proof does it show? Screenshots, test data, original numbers, a named author who clearly did the thing. Pages that show their work win categories where everyone else paraphrases.
A useful gut check: if a stranger read both pages, would they be able to tell yours was written by someone who had actually done it? If not, no amount of on-page tweaking will hold the position.
Step 5 — only now, consider a core update
Match your drop date against Google's published list of ranking updates. Two rules:
- The dates have to line up closely. A drop that started three weeks before an update was not caused by that update.
- Core updates move sites, not pages. If only one page fell, it was not the update.
And if it genuinely was one: there is no switch to flip. Google's own guidance has been consistent for years — there is nothing to "fix", the assessment of the whole site changed. What recovers sites is the unglamorous version of the same advice: fewer, better pages; cut or merge the thin ones; make the good ones unmistakably first-hand; and stop publishing things only to have something to publish.
What to actually do about it
Whatever the cause, the response is the same shape — and the order matters, because the top of this list takes minutes and the bottom takes days:
- Fix anything technical. Indexing, redirects, noindex, speed. Then request indexing.
- Fix the snippet before the page. If your click-through rate is far below what your position normally earns, a better title and meta description buys back traffic without moving a single rank.
- Close the coverage gap. Add the sections the winning page has and you do not. Put the direct answer in the first hundred words.
- Add proof. One real screenshot, one number from your own testing, one specific example beats another five hundred words of explanation.
- Update the date honestly. Change the content, then change the date — never the other way round.
- Re-submit and wait properly. Request indexing, then leave it alone for two to four weeks. Changing a page every three days makes it impossible to learn anything from the result.
Then measure. Export Search Console again in a month and compare the same page's position, impressions and CTR against what you wrote down today. That comparison — your change, your page, your numbers — teaches you more about your own site than any general SEO advice, including this article.
I run a content site, got tired of paying four subscriptions to find out what to write, and built this instead. Everything here is tested on my own traffic first.
Want this done for you?
Fix My SERP does everything in this guide automatically — finds the pages, ranks them by what you would gain, and gives you the list in order. Free, and it runs on your own hosting.
Start free What it doesRead next
Striking distance keywords: the traffic you already own
Positions 4 to 20 are where most small sites are quietly sitting on their next 40% of traffic. Here is how to find those keywords in a free Search Console export, work out which ones are worth the afternoon, and move them.
How to update old blog posts (a refresh checklist that works)
Refreshing a post that already ranks beats writing a new one almost every time. The catch is that most "refreshes" are date changes with a paragraph bolted on. Here is how to pick the right posts and what to actually change.
Turn your Search Console export into an action plan
Everyone exports the CSV. Almost nobody does anything with it. Here are the six things to pull out of that file, in order, and what each one tells you to do this week.