Google Algorithm Volatility Explained: What the Score Actually Measures

What a volatility score actually counts

Every volatility tracker works the same way in principle. Take a fixed set of keywords, record the top results for each one every day, and measure how much the results moved between one day and the next. The number that comes out is displacement, normalised onto a scale that means something to a reader.

Because this site runs its own sensor, I can describe the method rather than guess at somebody else’s. The volatility sensor here works like this:

• A fixed basket of 150 keywords across 15 categories. Why: Ten per category, spanning health, finance, travel, shopping, local services and others. The basket never changes, because swapping keywords would break comparability with every earlier reading.
• One snapshot per keyword per day, at the same hour. Why: Results shift through the day. Fixing the collection time removes that as a variable.
• The top 10 organic domains recorded for each keyword. Why: Positions beyond ten move constantly and add noise rather than signal.
• Each day compared against the previous day. Why: The score is how far each domain moved, plus a heavier weight for domains entering or leaving the top ten entirely, since that is a bigger event than shuffling from four to six.
• Normalised to a 0 to 10 scale. Why: Raw displacement figures mean nothing to a reader. The scale is set so a quiet day reads under 2 and a major shake-up reads above 8.

The full documentation, including current scope and known limits, is on the methodology page.

Why two trackers disagree on the same day

One tool says 8.4, another says 5.1, a third says the day was unremarkable. This confuses people into thinking one of them is broken. None of them is. There is no shared standard for measuring volatility, so every tracker is measuring a different thing and calling it the same name.

Variable	Why it changes the number	Measurement (what to check)
Keyword basket	A basket weighted to news moves more than one weighted to reference queries	Does the tool publish its keyword count and categories?
Basket size	Small baskets are noisier; large ones smooth real events away	Anything under about 100 keywords swings on individual results
Depth measured	Top 10 versus top 20 produces very different displacement	Deeper tracking almost always reports higher volatility
Device and country	Mobile and desktop results diverge, and so do markets	US desktop is the default for most tools, including this one
Normalisation	How raw displacement is mapped onto 0 to 10 is an editorial choice	One tool’s 7 is another tool’s 4
Collection time	Results change through the day during a rollout	Morning and evening snapshots can differ materially

The practical rule: compare a tracker against its own history, never against another tracker’s number. A reading of 6 means something if you know what that tool’s normal week looks like. It means nothing in isolation.

Reading the bands

Score	What it means	Measurement (what to do)
Under 2	Quiet. Ordinary daily shuffling	Nothing. Do not read anything into it
2 to 5	Normal. Google ships changes constantly	Nothing, unless your own data moved
5 to 8	Elevated. Something is happening somewhere	Check your own Search Console before concluding anything
Above 8	High. Usually a confirmed update or a large test	Check the update tracker, then your own data

Note what the right-hand column keeps saying. Every band ends at your own Search Console, because that is the only place where the question of whether your site moved can actually be answered.

The mistake this page exists to prevent

A high volatility reading is not evidence that your site was affected. It is evidence that some results moved for some keywords in some categories, most of which have nothing to do with you.

I see this pattern regularly: a site owner reads that volatility spiked, checks their traffic, finds a normal weekly dip, and concludes they were hit. Work begins. Two months later there is still nothing to recover from, because nothing happened.

The reverse error is more expensive. A site drops sharply on a day when volatility was quiet, and the owner concludes it cannot have been an algorithm change. Often it was: a narrow adjustment affecting one page type in one niche will not register at all in a 150-keyword basket spread across 15 categories. Low volatility is not evidence of your safety either.

Volatility answers one question only: did something change out there. Whether it changed for you is a separate question with a separate data source, and the drop analyzer answers it from your own export.

Why volatility tracking matters more than it used to

In 2021 Google confirmed eleven updates. In 2025 it confirmed four, while measured volatility stayed high. That divergence is the whole argument for watching volatility at all.

Five years ago, most meaningful ranking changes could be matched to an announcement. Today most cannot. October 2025 is the clearest example: Google confirmed nothing between the August 2025 spam update and the December 2025 core update, an 80-day gap, yet three separate waves of volatility hit during October, with third-party sensors peaking around 7.5 on 16 October.

For anyone whose traffic moved that month, the confirmed update list was useless and the volatility record was the only external evidence available. That is why the update tracker and the sensor sit alongside each other rather than one replacing the other.

A caveat about the 2025 readings

Worth stating because it affects how much weight to put on historical figures. In mid-September 2025 Google removed the ability to load 100 results per page. Every rank tracking tool had been using that parameter to collect data efficiently, and all of them had to recalibrate.

For several weeks afterwards, tool readings were noisier than usual and less comparable with earlier periods. Readings from late 2025 are directionally useful and less precise than a normal month. The same change also caused a wave of Search Console impression drops that had nothing to do with rankings, which is covered in the seven causes of a drop with no update.

Using volatility properly

• Use it to date something, not to detect it. Why: Start from your own drop date and look up what volatility was doing then. Starting from the volatility chart and hunting for a matching drop is how people convince themselves of things that did not happen.
• Check the category, not just the overall score. Why: An overall reading of 4 can hide health at 8 and travel at 2. If your vertical is the one moving, that is far more relevant than the headline number.
• Treat a spike as a prompt to look, never as a finding. Why: The finding is always in your own data. Volatility just tells you when it is worth checking.
• Do not act on volatility alone. Why: Changing a site in response to movement you have not confirmed affected you is how a healthy site acquires a real problem.
• Watch it during a confirmed rollout. Why: Core updates take 12 to 24 days and volatility usually stays elevated throughout. When it settles, the rollout is finishing and your data becomes reliable again.

If you run a local business, none of this applies in the way you would expect. Local rankings change with the searcher’s location, so a sensor tracking one fixed point misses almost all of it. That is covered in local search volatility.

Which tracker to watch, and what each one structurally cannot see, is compared in tracking tools compared.

If you would rather not check daily, the free alerts email you only when volatility crosses 5 without a confirmed update, or when Google confirms one. Two triggers, nothing else.

Frequently asked questions

What is Google algorithm volatility?

A measure of how much search results moved across a tracked set of keywords compared with the previous day, usually expressed on a 0 to 10 scale. It describes movement across the web, not on any individual site.

Does high volatility mean my site was hit?

No. It means some results moved for some keywords, most of which have nothing to do with your site. Whether you were affected can only be answered from your own Search Console data.

Why do volatility tools show different numbers?

Because there is no shared standard. Tools differ in keyword basket, basket size, ranking depth, device, country and how raw displacement is mapped onto a scale. Compare a tool against its own history rather than against another tool.

What is a normal volatility score?

Between 2 and 5 on most scales. Under 2 is quiet, 5 to 8 is elevated and worth checking your own data against, and above 8 usually indicates a confirmed update or a large test.

Can my site drop when volatility is low?

Yes, and it happens often. A narrow adjustment affecting one page type in one niche will not register in a broad keyword basket. Low volatility is not evidence that nothing happened to you.

Should I do anything when volatility spikes?

Check your own Search Console, and nothing else. Do not change the site in response to volatility you have not confirmed affected you. If a rollout is running, changes made mid-rollout also make the result impossible to attribute.