Google’s ranking volatility no longer means what it used to

Search Engine Roundtable flagged another unconfirmed Google ranking update over the weekend of July 24th. Third-party trackers spiked. WebmasterWorld filled with the usual: 60-70% traffic drops in Brazil, ghost-town weekends in the UK, zero conversions on Google Ads for the first time since launch. All the familiar signals of a Google shakeup.

And Barry Schwartz wrote something in that piece that I haven't been able to stop thinking about. "The sad thing is, these ranking movements continue to matter less and less for publishers because Google is sending less and less traffic to publishers."

That's from the person who has covered every Google update for twenty-plus years. When Barry says the updates matter less, that's not a hot take. That's the guy who invented the beat quietly telling you the beat is changing shape.

The interesting thing about this particular update isn't the update. It's what the industry response is showing us about where SEO measurement is actually broken.

The volatility charts are measuring the wrong game

Look at what the tracking tools measure. AWR, DataForSEO, Wincher, Mozcast, AccuRanker Grump, SEMRush — they all sample keywords, watch position changes, and report the delta. That was a fine methodology when the top ten blue links accounted for most of the traffic reaching a website. It's a progressively worse methodology every quarter that AI Overviews, AI Mode, and Gemini answer more of the query before the click happens.

Position three on a query that now resolves inside an AI Overview isn't the same position three it was two years ago. The tool still reports it as position three. The volatility chart still lights up when it moves to position four. But the click-through curve underneath that position has been quietly collapsing, and no rank tracker is showing you that.

So when a chart spikes, you're seeing that Google shuffled the deck. You're not seeing whether the deck matters.

The comments underneath are the actual signal

Read the WebmasterWorld chatter Barry pulled. "60% to 70% drop in Brazil." "27% of normal traffic." "Zero conversions on Google Ads." These aren't ranking complaints. They're volume complaints. People aren't saying "I moved from three to seven" — they're saying "traffic collapsed."

That distinction matters. Historically, ranking movement and traffic movement tracked each other closely enough that people used the terms interchangeably. They're decoupling now. You can hold rankings and lose traffic. You can lose rankings and hold traffic. You can lose both simultaneously for reasons that have nothing to do with the ranking algorithm and everything to do with how much of the query is being answered before the SERP renders.

Which means the diagnostic question SEOs have been trained to ask — *did I lose rankings?* — is answering the wrong problem. The right question is *what portion of my query volume is still producing a click, and is that portion shrinking?*

Almost nobody is set up to answer that.

What Google actually said this week

Alphabet's Q2 numbers landed the same week as this ranking chatter. Search & Other revenue up 17% year-on-year to $63.27bn. Sundar Pichai crediting "popular AI features" for query growth. Nick Fox saying Search hit its highest usage on record during a World Cup match.

When the number matters to Google, Google gives you the number. When it doesn't, you get a superlative.

Notice what's precise and what's vague. Revenue: precise to the decimal. Growth rate: exact percentages, quarter-on-quarter comparisons. Click volume to the open web: "billions" — daily in one statement, weekly in another, with no reconciliation between the two.

When the number matters to Google, Google gives you the number. When it doesn't, you get a superlative.

Liz Reid's "quality clicks" framing from last August is the tell. Organic click volume is "relatively stable" but "quality" has increased. The definitions are Google's. The measurement is Google's. The comparison to prior periods is Google's. And the third-party reports of traffic drops are dismissed as "flawed methods and isolated cases."

You don't need to be conspiratorial about this. You just need to notice that the entity with the data isn't sharing the data, and the entity without the data is being told to trust the summary.

The measurement problem is the actual problem

I've written about this before but this week made it concrete. The ranking volatility trackers, the traffic drops on client dashboards, and Google's "billions of clicks" reassurances are all measuring different things — and none of them are measuring what a business actually needs to know.

What a business needs to know: how many people in my target market saw my brand referenced this week, in any surface — organic result, AI Overview, AI Mode citation, ChatGPT answer, Gemini response, Reddit thread that got surfaced — and how many of them acted on it.

Nobody sells that dashboard. The rank trackers can't see AI surfaces properly. GA4 can't distinguish an AI-referred click from any other referral once the user lands. Brand-monitoring tools see mentions but not query context. AI citation trackers sample prompts but can't tell you frequency-weighted exposure. Everyone has one shard of the picture.

So when a Google update hits and rankings move, the honest answer to *"how bad is this?"* is *"we don't know, because we haven't been measuring the thing that matters for eighteen months."*

What this actually means for the reader

If you run a site and you saw movement over the weekend, the temptation is to react to the ranking chart. Don't. Here's what to do instead.

Look at your traffic against your ranking movement across the last six months, not the last week. If those two lines are diverging — rankings roughly stable but traffic bleeding, or rankings bouncing but traffic flat — you're seeing the AI compression, not a specific update. That's a strategic problem, not a tactical one, and rewriting title tags won't fix it.

Look at your branded search volume. If branded search is holding or growing while non-branded traffic drops, your brand is doing the work that Google used to do for you, and you should invest in that side harder. If branded is also dropping, that's the more serious signal — it means the AI layer isn't putting you in front of the audiences that would have converted to branded searches downstream.

And stop watching the volatility trackers as if they're a market indicator. They're an interesting weather report from a climate that's changing faster than the instruments can measure. Useful, but not what you optimise against.

The close

Barry's throwaway line — that these updates matter less because Google sends less traffic — is the whole story compressed into a sentence. The SEO industry built its rituals, its tools, and its client conversations around a world where rank movement equalled traffic movement. That equivalence is dissolving in real time, and the update charts are the last honest witnesses to a game whose stakes have quietly moved elsewhere.

The businesses that will handle the next three years well aren't the ones watching volatility spikes. They're the ones building brand into the discovery layer that AI systems actually cite, and accepting that a portion of their old traffic isn't coming back regardless of what the ranking tools say.

That's the loop. And most of the industry is still watching the wrong dial.