Singapore
Reading the November 2026 Core Update's Ranking Volatility

Reading the November 2026 Core Update's Ranking Volatility

As of November 2026, the honest answer to "did the November 2026 core update hit my rankings" is that the movement is real but unevenly measured. The update began rolling out on November 10 at around 1 pm ET, Google said the rollout may take up to 2 weeks, and the trackers that usually settle the argument are reading very different numbers: Similarweb's desktop index sat at 24 on November 16 against a last-month average of 51, while MozCast showed 123° for the same Monday against an uneventful day of about 70°F. Any team weighing best seo singapore shortlists this month should read that gap as a measurement problem before reading it as a ranking problem, because the two indexes are not answering the same question.

Key takeaways

  • The rollout began on November 10 at around 1 pm ET, and Google said it may take up to 2 weeks to complete.
  • The aggregate tool reading lifted on Saturday, November 14, and the search chatter points to the update taking effect 3-4 days after the announcement.
  • One index number is never the whole story: MozCast tracks 10,000 keywords across 20 industry categories and 5 major US cities, while AccuRanker averages rank changes across about 1,000,000 high-volume keywords.
  • Decide with per-keyword history across days, compared against competitors, before changing a single page.

What happened between November 10 and November 16

The sequence is short and worth keeping in order. Search Engine Roundtable recorded that the November 2026 core update started rolling out on November 10 at around 1 pm ET, with Google describing it as a regular update designed to better surface relevant, satisfying content for searchers. Google also said the rollout may take up to 2 weeks to complete, which means the final shape of the results was never going to be visible in the first few days.

What surprised the industry was the volume of the reaction. The report notes that chatter around this update was more muted than most, and it points at the third-party tracking tools as part of the explanation. The tools, taken together, still showed more volatility on November 14 than on other days, and the lift in chatter over that weekend is what leads the report to conclude the update may have kicked in 3-4 days after its announcement. Google Search Console's performance reporting was also lagging, which removes the one first-party view a site owner trusts most.

The rollout clock

Two weeks of rollout changes what a fair reading looks like. A dip recorded on the second day of a rollout is a partial signal: only part of the index has been rewritten, so a keyword can move twice in a fortnight for reasons that have nothing to do with the page. The practical move is to fix the reading window now, note the starting position for each keyword, and refuse to draw a conclusion until the rollout window closes.

Why the chatter arrived late

People notice ranking changes when they look, and most look on a Monday or over a weekend. The report's own timeline has chatter lifting on the Saturday, which is consistent with an update that takes a few days to reach enough of the index to be visible in ordinary site data. A quiet Tuesday and Wednesday is not evidence the update missed your market; it is evidence that the sample was still small.

The tools disagree because they sample different things

The comparison table below is the part most commentary leaves out. Each index is built from a different set of queries, read on a different rhythm, and scored by a different formula.

IndexWhat it samplesReading on November 16
Similarweb SERP Seismometer, desktopDomains and keywords it monitors daily, desktop and mobile scored separately24 on November 16, against a last-month average of 51
Similarweb SERP Seismometer, mobileThe same panel, reported on its own scale32 on November 16
MozCast10,000 keywords across 20 industry categories and 5 major US cities, page one compared with the previous day123°, where an uneventful day is about 70°F
AccuRanker Google GrumpAbout 1,000,000 high-volume keywords, the top 30 results compared with the previous day's rankingsAn average count of rank changes, not a temperature

Read the first three rows together and the apparent contradiction dissolves. Similarweb builds its score from a fixed panel and reports desktop and mobile separately; MozCast compares page one of a hand-picked set against the day before and expresses the result as a temperature. A score of 24 and a temperature of 123° can describe the same Monday because one is a volatility level relative to that panel's own history and the other is a change count dressed as weather.

Two gauges on a desk show very different readings beside a mid-November calendar.

Reading a low score correctly

A desktop score of 24 sits below the last-month average of 51 and below the three-month average of 41, which the index itself labels stable ground. The panel looked calmer that day than it had through the preceding month, though the mobile score of 32 tells a slightly different story for phone results. For a Singapore business where most local searches happen on a phone, the mobile number belongs beside the desktop number in the report.

Why a hot MozCast and a cool Similarweb can both be true

AccuRanker makes the mechanism plain: it loops the top 30 rankings for each keyword and compares them with the previous day, then divides the total difference by the number of results. MozCast does the same kind of comparison on page one of its 10,000 keywords. Similarweb keeps both desktop and mobile series and reports them against its own monthly and three-month averages. Four samples, four cadences, one Monday. A flat-looking index is therefore not evidence that nothing moved in your category; it is evidence that the category was not the loudest part of that sample.

A three-signal test before you touch anything

The first two weeks of a rollout are the worst possible moment for a rebuild, and the best moment for evidence. Run these three checks in order, and do not start with the site.

  1. Re-read the same keyword on two separate days and compare the position, the URL that ranks, and the search result layout. A drop that reverses on the second read was noise.
  2. Put your movement beside the movement of two competitors for the same query. If everyone moved, the cause is upstream of your pages.
  3. Check the technical floor: broken links, page speed, index coverage and whether the ranking URL is the one you meant to rank.

Signal one: does the drop survive a second reading?

A single reading is a photograph of a list that changes hourly during a rollout. Two readings on different days turn it into a trend with a direction. Write both down with the date, the device and the market, because a mobile position in Singapore and a desktop position in the United States are not the same measurement and should never be averaged into one line.

Signal two: did competitors move too?

This is the check that separates an algorithm-wide shift from a site-specific problem, and it is the reason any local seo singapore report should always carry competitor activity beside your own position. When three sites trade places on the same day, no amount of on-page work explains it. When only your page falls while the same three sites hold, the question moves to your content, your intent match or your technical setup.

Signal three: is the page still technically sound?

Technical faults produce the same symptom as an update: a page that stops ranking. A broken redirect, a blocked resource or a slow template can take a page out of the results without any algorithm change at all. Check the floor before you rewrite the content, or you will spend the fortnight fixing prose on a page that never loaded properly.

How to read a rank report during a rollout

The report format decides what you can see. A monthly average hides the curve; a daily series shows it.

Daily, weekly or monthly

Pick the cadence to match the question. Daily suits the fortnight of a rollout, weekly suits the month after it, and a monthly summary is the right shape for a management update. One report cannot serve all three purposes, and a single monthly average cannot tell you whether a drop happened in one day or over four weeks.

The single-day trap

The most expensive mistake in a volatile week is treating one day as a verdict. A keyword that leaves the top group on a Tuesday and returns on a Thursday has not been penalised; it has been reordered while the index was being rewritten. Record the day, keep the URL you saw, and wait for the pattern.

A marketer sits back from a laptop beside a printed daily ranking chart, choosing to wait.

Where AI carries the first pass, and where a person still signs off

The work splits cleanly, and the split is worth writing down before the next update. A model can take the keyword export, group the terms by the page that serves them, flag which groups moved together and draft the summary paragraph a manager reads. That is collection and pattern work over a fixed data set, and it is fast.

What the model cannot do is decide whether the page still answers the query. It does not know that the service page now leads with a price the sales team withdrew, or that the copy drifted away from what the searcher actually wants. That judgement needs the person who owns the page, and it should be a named review point on the calendar, not a comment in a chat thread. The same boundary applies to any local seo singapore programme: AI can assemble the evidence, but a human signs off on the content decision.

The Topkee view: diagnose per keyword before blaming the algorithm

Topkee's position, as set out in its own material, is that a ranking drop is a diagnosis before it is an explanation. Its Rank Tracker takes the keywords and the target country or search engine and returns current positions with historical data, which the material frames as the way to quickly detect the cause of ranking fluctuations; reports can run daily, weekly or monthly and list current rank, the historical trend and competitors' SEO activity. The same material attributes ranking drops to three operational causes: intensified competition, content mismatch and technical problems, and pairs continued monitoring with a site audit covering 20+ indicators, from load speed and mobile compatibility to content originality, backlink quality and user experience, followed by periodic automated re-checks.

That framework describes diagnosis, not prediction, and the honest use of it is narrow: it tells a team whether the movement looks industry-wide or local to the site, which is the question that decides whether the next step is content, competition or engineering. Topkee Pte. Ltd., the group's Singapore entity, keeps its local SEO work on local keyword research, consistent business listings and locally written content, with rank tracking kept to the country and market in question.

What to ask an agency in a volatile week

A volatile week is a good time to test how any best seo singapore contender reads data, because the answer cannot be a single number. Ask for the keyword list, the market, the device and the dates behind any claim about the update, then ask to see two readings of the same keyword. Ask how they separate an industry-wide shift from a site-specific fault, and which of the three causes they would rule out first.

Be equally careful with the opposite pitch. A guaranteed first-page promise is a commercial commitment about positions, not a forecast about an algorithm, and a top ranking that does not convert still costs money. During a rollout, a team that shows you the work behind the number is worth more than a team that promises the number.

What to watch next quarter

Three things are worth a calendar note. First, whether the per-keyword positions you recorded at the start of the rollout hold once Google's window closes, which is the only fair test of whether the movement was the update or your own change. Second, whether the two index families settle into the same story, because a month of disagreement between a panel-based score and a page-one temperature is a signal about measurement, not about your site. Third, whether the same pages keep their position in AI answers, since a page that holds its classic rank and disappears from an answer has a different problem from a page that fell, and that check belongs in the same best seo singapore review as the rank line. The answer engine optimisation practices article covers how that second surface is read. For background on the search-side changes that landed just before this update, the finished November spam update report sets out what Google actually shipped.

Colleagues review a quarterly wall planner with three pinned markers and a phone on the table.

FAQ

Did the November 2026 core update finish on November 16?

No. Google said the rollout may take up to 2 weeks from the November 10 start, so a reading taken on November 16 sits inside the window. Treat positions recorded before it closes as provisional.

Why did my rankings barely move when other sites report big drops?

Because the trackers are not measuring your pages. MozCast reads a hand-picked set of 10,000 keywords across 20 industry categories and 5 major US cities, AccuRanker works across about 1,000,000 high-volume keywords, and Similarweb reports its own panel with desktop and mobile scored separately. Your category may simply not be the loudest part of any of those samples.

Is a lower Similarweb score better or worse?

Lower means calmer. A desktop score of 24 on November 16, against a last-month average of 51 and a three-month average of 41, reads as stable ground in the index's own terms, and the mobile score of 32 points the same way. A high score would mean more movement, not better rankings.

Should I rewrite content in the middle of a rollout?

Not on the strength of one reading. Fix technical faults first, because they produce the same symptom and are measurable, then record the keyword history and the competitor comparison. Content changes made before the window closes cannot be separated from the update when you review them later.

How long before I know whether a drop is mine or the algorithm's?

Long enough for two readings on different days and a competitor comparison on the same query, and ideally until the rollout window closes. A local seo singapore report built on daily or weekly readings will show the shape of the movement; a monthly average will not.

Sources

  1. "Google November 2026 Core Update Ranking Volatility", Search Engine Roundtable, 2026-11-16. https://www.seroundtable.com/google-november-2026-core-update-ranking-volatility-42543.html
  2. "SERP Seismometer - Check SERP Volatility in Google Updates", Similarweb, 2026-11-16. https://www.similarweb.com/serp/
  3. "The Google Algorithm Weather Report", Moz, 2026-11-16. https://moz.com/mozcast
  4. "Google Grump Rating", AccuRanker, 2026-11-16. https://www.accuranker.com/grump/
Category:
Share to:
Date: 2026-11-17
Terry Wong

Article Author

Terry Wong

Website Production Manager