Fixing High Ranking Volatility
Some pages never settle. One week a page sits at position 4, the next at 11, then back to 6, while the competitors around it barely move. There is no single update to recover from here: the swinging is the problem.
It usually means Google is uncertain which of your pages to show, whether your page is the right kind of answer, or whether it can reliably fetch the page at all. This guide helps you separate genuine instability from measurement noise, find which cause applies, and make changes that give Google a clearer, steadier picture.
First, is it really volatility?
A good share of "volatility" is how rankings are measured rather than how they behave. Before changing anything, rule out these artefacts:
- Average position is a blend. Search Console averages your position across every query, device and location where the page appeared. If the page starts appearing for an extra low-ranking query, its average worsens even though nothing moved for your main query.
- Results differ by place and device. Local intent, language and device all change the results. A rank tracker checking from one city will not match what customers in another town see.
- Result features are counted differently. Some tools count a featured snippet, a video carousel or a "People also ask" box as a position and others do not. A new feature above you can look like a two-place drop.
- Sampling. A tracker that checks once a week can catch a page on an unusual day. Daily data smooths that out.
- Low volume. For queries with few impressions, a handful of searches can move the average a long way.
The quickest test: in Search Console, filter to one exact query and one country, choose one device, and look at the daily position line. If that line is steady, you have measurement noise, not volatility.
What causes real volatility
| Cause | What you will see | Where to confirm it |
|---|---|---|
| Keyword cannibalisation | Two or more of your URLs appear for the same query on different days, each ranking lower than one page would | Search Console, filter by query, then the Pages tab |
| Unstable indexing | The page drops out entirely for a few days, then returns | URL Inspection, Page indexing report, server logs |
| Canonical flapping | Google alternates between your chosen URL and a duplicate as canonical | URL Inspection: user-declared vs Google-selected canonical |
| Mixed intent | The query's results mix guides, shops and tools, and Google keeps testing which type works | Look at the results for the query on several days |
| Freshness demand | New articles push you down briefly after news or seasonal events, then fade | Dates of the pages ranking above you when you drop |
| Borderline relevance | The page answers part of the query, so it trades places with pages that answer other parts | Compare what your page covers with the stable top three |
| A new page | A recently published page moves a lot in its first weeks | Publication date; usually settles on its own |
Most persistent swinging on small business sites comes from the first three. They are also the most fixable, because they are about your own site's signals rather than Google's opinion of the content.
The fix, step by step
Step 1
Check the swings are real
Pick the query you care about and, in the Search Console Performance report, filter to that exact query, one country and one device, then view daily average position over three months. Compare with the Algorithm Impact monitor, which shows daily ranking movement alongside tracked Google updates. If the filtered line is steady, the swings came from blended averages or tracker sampling, and no change to the page is needed.
Step 2
Look for your own URLs taking turns
With the query filter still applied, open the Pages tab. If two or more of your URLs have impressions for the same query, switch the chart to daily and add a page filter for each in turn to see whether their appearances alternate. Alternating URLs are the clearest sign of cannibalisation: Google cannot decide which page is your answer, so each ranks lower and less steadily than one would.
Step 3
Check Google can fetch the page every time
Run the page through URL Inspection and compare the user-declared
and Google-selected canonical. Then search your server logs for Googlebot requests to the URL and look
for 5xx responses, 429 rate limiting or timeouts. Firewalls, bot protection
and overloaded hosting can refuse some Googlebot visits but not others, which makes a page drop out for
days at a time and then return as if nothing happened.
Step 4
Match the page to the dominant intent
Search the query on several different days and note the type of page in the top five each time: guide, product page, category, tool, video or forum. If the mix keeps changing, Google is still testing intent. Make sure your page clearly serves the type that appears most often, and put the direct answer near the top so the page is useful whichever angle the searcher had in mind.
Step 5
Consolidate and strengthen the chosen page
Where cannibalisation was found, choose the stronger URL, move the useful material
from the weaker page into it, and 301 redirect the weaker URL to the chosen one. Fix any
duplicate versions so they point to one canonical. Then fill the gaps you found when comparing with the
steady top three, so the page answers the whole query rather than one slice of it, and keep the title
focused on that query.
Step 6
Point internal links at one clear target
Crawl the site and list every internal link whose anchor text matches the query. Update links that point at the old or competing page so they point at the chosen URL, and add a link from your most relevant strong pages. Check the chosen URL is reachable from your main navigation or hub pages, and update the sitemap to list only the canonical address. Consistent internal signals make Google's choice easier and steadier.
Step 7
Judge stability over a longer window
After making changes, wait for Google to recrawl, then compare the spread of daily positions over four to eight weeks with the same period before. A narrower range matters more than a better single day. Use the Algorithm Impact monitor to check whether any remaining swings coincide with tracked updates, in which case they reflect wider movement rather than a problem specific to your own site.
Resolving cannibalisation without losing traffic
Not every pair of pages ranking for the same query is a problem. A category page and a buying guide can both deserve to appear, and Google sometimes shows two results from one site. It becomes a problem when the pages alternate and both rank worse than you would expect. Decide which case you have:
- Same intent, same content: merge into one page and redirect the other.
- Different intent, overlapping wording: keep both, but sharpen each so its purpose is obvious from the title, headings and opening paragraph, and link between them.
- Old and new version of the same article: keep the better URL, bring the best of both into it, redirect the other.
When merging, redirect at server level so the old address sends a permanent redirect in one hop. On Apache:
Redirect 301 /blog/boiler-service-cost-2024/ https://www.example.co.uk/guides/boiler-service-cost/On nginx:
location = /blog/boiler-service-cost-2024/ {
return 301 https://www.example.co.uk/guides/boiler-service-cost/;
}Intermittent fetching problems
A page that vanishes for a few days and then returns at its usual position often has nothing wrong with its content. Google simply could not fetch it reliably. Common culprits:
- Bot protection that challenges some Googlebot requests, especially after a burst of crawling. Verified Googlebot traffic should be allowed; verify it with a reverse DNS lookup rather than trusting the user agent.
- Rate limits returning
429or503to crawlers during busy periods. - Shared hosting timing out under load, which shows as server errors in the Crawl stats report under Settings in Search Console.
- Geo-blocking that blocks traffic from outside the UK, which can include Googlebot, as it often crawls from the United States.
- Rendering failures where the main content depends on a script that sometimes errors, so Google indexes an empty shell on some crawls.
To check the logs for failed Googlebot requests to one page:
grep "Googlebot" access.log | grep "/guides/boiler-service-cost/" | awk '{print $9}' | sort | uniq -c | sort -rn | headAny status other than 200 or 304 in that list is worth investigating.
When volatility is normal
Some movement is simply how search works, and chasing it wastes effort:
- New pages often move around for several weeks while Google gathers signals about them.
- News and seasonal queries reshuffle whenever fresh articles appear. A steady guide can be pushed down by news coverage for a few days and return afterwards.
- Update rollouts cause movement across whole niches. Google's announced updates, and unconfirmed changes reported by third-party trackers, can move many sites at once.
- Positions 8 to 15 are crowded and naturally less stable than the top three, because many pages are close in quality.
If a page has swung within a narrow band for months without trending down, and your traffic is stable, the practical answer may be to improve the page for readers and stop watching daily positions.
1. Why do my rankings go up and down every week?
The most common reasons are measurement noise, keyword cannibalisation and unreliable fetching. Average position blends many queries, devices and places, so it moves even when your main ranking does not. If two of your pages compete for the same query, Google may alternate between them. And if your server or firewall sometimes refuses Googlebot, the page can drop out and return. Filter Search Console to one query, one country and one device first to see which you have.
2. Is ranking volatility a sign of a Google penalty?
Rarely. A penalty in the usual sense is a manual action, which appears in Search Console and causes a sustained drop rather than weekly swings. Algorithmic updates also tend to cause a shift that then holds. Persistent back-and-forth movement is more often Google being unsure which page to show, or whether yours is the right type of answer. That uncertainty is fixable through clearer signals, which is a much better position to be in than a penalty.
3. How do I find keyword cannibalisation in Search Console?
Open the Performance report, add a query filter for the exact search term, and then open the Pages tab. If more than one of your URLs has impressions for that query, you have overlap. Switch to a daily view and filter to each page in turn to see whether they appear on alternating days. Pages that alternate, each ranking lower than you would expect, are the case worth fixing, usually by merging them or sharpening each page's purpose.
4. Should I merge pages that compete for the same keyword?
Merge them when they serve the same intent and cover similar ground. Move the best material into the stronger URL, redirect the weaker one with a permanent 301 redirect, and update internal links to point to the survivor. Keep them separate when they genuinely serve different needs, such as a product category and a buying guide, but make each page's purpose obvious from its title and opening so Google and readers can tell them apart easily.
5. Can hosting or a firewall cause rankings to fluctuate?
Yes. If your server times out under load, or bot protection challenges some Googlebot requests, Google may fail to fetch the page on some visits. Repeated failures can lead to the page dropping out of results for a while and returning when fetching succeeds again. Check the Crawl stats report in Search Console for server errors, and search your access logs for Googlebot requests to the affected page that returned anything other than a 200 or 304 status.
6. How long does it take for rankings to stabilise after changes?
It depends on how often Google recrawls the pages involved. Redirects and canonical changes on well-linked pages are often picked up within days or a few weeks, while less visited pages take longer. Judge stability over four to eight weeks rather than day to day, and compare the range of positions rather than the best or worst single day. A narrower range, even at a similar average, is the sign that your changes have worked.
7. Do unconfirmed Google updates cause volatility?
They can. Google makes many changes it does not announce, and third-party trackers often report days of unusual movement across the results. Those changes can move your pages along with everyone else's. The difference is that they affect whole niches, not just your site, and they come and go. If your page swings while competitors stay steady for months, an unconfirmed update is an unlikely explanation, and the causes on your own site are worth checking first.
The Algorithm Impact monitor lines up your daily ranking movements with tracked Google updates, helps separate update effects from technical faults, content decay and competitor gains, tracks your recovery and suggests actions suited to the type of update.
Open Algorithm Impact Fixes →