July 2025 Google Search Observations

[Martin Ice Web]

And like every time, google lies about he start of this core update. It definitly started on friday.
And again it rewards old fashion, low content, bad html websites. Even websites with no pics or pictures smaller than 300 width are rewarded.

[edited by: not2easy at 11:55 am (utc) on Jul 1, 2025]
[edit reason] New month, new thread [/edit]

[mosxu]

@EG

Are you actually serious? Who is stay22?

On a different note I am using G about 20% since ChatGPT and Grok!

It is happening!

[Whitey]

A blog post from Stay22.com, an affiliate platform for travel sites, has an interesting take on the June Core Update's impact:

"Traffic Is Up for Travel Bloggers - June 2025 Google Update"

@EditorialGuy - i think the devil will be in the detail, rather than an "overall" picture that averages out the lift. There will be winners and losers among the 4000+ bloggers. This affiliate network depends on revenue from their bloggers, so anything they can do to inject some enthusiasm is in their interest. Their business model will be under pressure, unfortunately.

The reality, I feel, is that producing text based content is on a trajectory to nowhere in the face of AI. It has to be multimodal these days, top quality, shareable and structured for citation retrieval, unique and delivered with brand reputation behind it. I doubt if a substantial proportion of the 4000+ bloggers will be responding to those requirements.

Which begs the question, why did Google let a whole lot of site's have improved visibility? One answer that springs to mind is that participation in content production via the HCU was gutted, stalling incentives across the internet pool. There's not a lot left to be removed from the live index, Google has it all.

AI eats AI then produces AI and then loop the loop.

[Conro]

@Whitey Because many sites did not deserve to be hit by hcu, but a machine does not understand it, so they will have improved it To better recognize real sites and not spam ones created or pumped with AI to make money and that's it. Also, if you train AI with the web, you need variety, sites that specialize in a topic, all stuff that Google was annihilating. Google then leaving too much space to brands and forums will have realized that it was not a good idea or at least could not continue like this in the long term, also because in my opinion they had a collapse in advertising revenue from the advertising banners of older websites

[Martin Ice Web]

Chicken has left the house again! Big drop today. Zero user engagment. As i see in the logs it started yesterday at sharp 8pm.

[Micha]

Traffic remains low and rankings continue to improve, which is a strange contrast. According to Semrush, SERP volatility has increased again today. So Google is still playing its games. However, there is no discernible pattern.

[Martin Ice Web]

@micha, we lost 50% over night. Complete rubbish traffic.

[Micha]

@Martin Yes, 99 percent of Google traffic is currently nonsense. Google has once again chosen the perfect moment for this. Summer vacation and updates are too much for any small business.

[Martin Ice Web]

wow, worst friday ever so far. Must be a new update or the core in´t over yet.

[Whitey]

The tools are continuing to show volatility. I wonder what’s going on.
[seroundtable.com...]

[Micha]

I agree, I've never experienced a day like this before. Google is silent, even though there's definitely an update in progress.

[Whitey]

ChatGPT:

Why the June 2025 Core Update Volatility Is Still Ongoing (and for How Long)

TL;DNR:

Although Google declared the June 2025 Core Update finished on July 17, ranking volatility continues and may persist for another 1–3 weeks.

Here’s why:

•Google’s systems adjust in phases — helpful content, E-E-A-T, AI/RankEmbed, and spam filters all settle at different times.

•Re-crawling and re-indexing delays mean sites aren’t re-evaluated all at once — especially lower crawl-budget domains.

•User behavior signals (clicks, bounce, dwell time) feed back into ranking tweaks post-rollout (e.g., via NavBoost).

•Tools like Semrush and Mozcast are still showing elevated turbulence.

•Historically, updates take 2–3 weeks to settle after rollout ends (e.g. March 2024, Sept 2022, June 2021 updates).

Sources: SEO tool dashboards + analysts like Glenn Gabe, Barry Schwartz, Marie Haynes

[JeepersCreepers1]

Traffic lost almost 50%. Without any big changes for long time and SEO score based almost 100.

Some older websites with many internal links and not clear UX moved up in SERP. That's fact.

Maybe it's chance for small bad website and slap face for websites with strong branded keywords.

Or maybe something gone wrong with that update.

Finding an answer is not easy now. Maybe for 1 - 2 month, yes.

Time will tell.

[EditorialGuy]

@EditorialGuy - i think the devil will be in the detail, rather than an "overall" picture that averages out the lift. There will be winners and losers among the 4000+ bloggers.

Of course. That's always the case, but it doesn't mean data points should be ignored.

[RedBar]

@MIW ... Yep, traffic way down today plus, when using Statcounter, so far 34% from Huawei, Singapore, I'll wait for my server logs later to view the real metrics.

[mosxu]

Search is dead!

Seriously in two years time you will have Gemini with a max 20% market share!

It is all over folks!

[nordland]

Search will never die. People just search on different platforms and in different formats.

[Micha]

@mosxus It's not the search engine, but the websites and AI that learn from themselves and become even more stupid as a result.

Just a question: Since Google hasn't said anything about what's going on, could it mean that something went wrong with the core update and they're now trying to fix it? Sure, there are always fluctuations after a core update, but they're not usually as severe as they are at the moment.

[christianz]

Search will never die. People just search on different platforms and in different formats.

The AI Overview/ChatGPT type of serching is non-viable model. Because there is no value passed to the sources of information - websites. They need to figure out a way how to pay publishers for:

1) The scraping of content for training
2) Agents visiting websites and fetching info real time

I am witnessing both and in large volume. Which is why I completely blocked ChatGPT for both scenarios, subject to change once Sam Altman agrees to send me some money.

[ichthyous]

I also saw traffic trending down the entire week, and yesterday it was much lower. Search was -19%, USA and UK both down 19-20%. UK traffic was much lower all week and traffic to my home page is at -50% for some reason. Today it looks like a more normal traffic pattern is kicking in.

@christianZ and the others out there blocking AI bots completely. All I can say is that I totally understand the motivation, but the final outcome is far from being determined. We are still very early in the AI race and there will be a lot more competition. I am seeing more and more traffic from AI platforms...is it actionable? Not yet I think. I do think that we really don't have a choice but to stay in the game and have our names and our sites mentioned in the AIO results, or in ChatGPT or wherever we can. More and more people are shifting to this kind of search and they aren't thinking twice about abandoning the old Google search.

This mostly impacts informational sites that make a living from running ads, the rest of us who are not dependent on informational content as much are not being affected as badly. My article traffic is way down, but to be honest the people who came to my site to read the articles really never resulted in any sales (I don't run ads so those pages were never monetized). I added that content simply for the edification of visitors, to add more authority to the site, and to see if I could siphon some of them off and convert them to paying customers (answer: not really).

The articles are still top destinations to the site despite losing almost half the visits. The real problem is that my actual category pages on my site, where the 'products' reside have steadily lost traffic as AIO have been added, and even before as more widgets were continually added.

The bottom line with AI though, is that if the AI platforms end up driving all the legitimate / unique content creation out of the market and into bankruptcy, who will provide these datasets with fresh content? I'm not stating anything new here...but it appears that many sites are being driven under faster than anyone expected. What will be left is AI content feeding on AI content.

One concrete example...last year I made a bid for a website that was going under. The site had a treasure trove of unique articles written by actual writers in the field. While speaking with the current owners they brought up over and over how the new owner could jettison all of the expensive writers and just use AI to write the articles themselves, thereby cutting one of the main expenses. Of course, he would have told me anything to sell the site at that point, but I thought it was telling. They had moved on to another site and the owners use AI to write everything themselves now.

[ichthyous]

Some observations from GSC:

I ran a report comparing traffic from the last 28 days to the same period last year. I am seeing much higher impressions but the CTR is half or less for many high traffic searches. Some have drops in the CTR of two thirds.

Meanwhile, there is an increase in CTR for some searches, but mostly more obscure ones or mismatches with my site content. Overall my traffic is higher, but the CTR is abysmal compared to last year. New inquiries slowed dramatically starting in April and are at tiny trickle now...commensurate with past periods of recession, like 2008/2009 or 2020. I don't think the drop is all Google, AIO, etc...but the combined impact is huge. I'm planning for a very poor 2025 in general, thankfully helped by an extremely strong start to the year from Jan - March.

[Micha]

Today was the first day with normal traffic, the Discover traffic from the news site is back, and the shop even had quite good sales from Google buyers today.
I hope it doesn't drop again tomorrow.

[christianz]

@ichthyous Your site(s) are primarily e-com. For you blocking ChatGPT would not make sense.

My site is 100% information and it is always changing and updating. For me it is opposite - allowing ChatGPT to act as some kind of middleman between audience and my site (that takes my ad revenue) makes no sense.

[christianz]

When Google stops abusing its monopoly status and allows webmasters to disable AIOs for their sites (right now it is tied to featured snippets), I will block AIOs too - not being featured in any AIOs will increase my traffic.

Why? Because when I get featured in AIO the answer is actually correct and good, there is no need to go to my site (nobody clicks on the citations anyway). When I am not featured in AIO, information may be incorrect and taken from low integrity source and users who suspect it is incorrect will look at the actual web results and maybe find my site there.

This is why Google prevented webmasters from having any way to opt out of AIOs. Because opting out is good for webmasters and bad for Google's plans (of having all traffic and ad revenue to themselves).

[Whitey]

Our core update gains are reversing in the last 24hrs. Down around 50%.

[EditorialGuy]

@Whitey: Our Google pageviews were down 31% on Saturday, while other sources of traffic were mostly up. On the whole, we gained nicely from the update--or would have if the Google Trampoline Effect hadn't returned

[Micha]

When Google stops abusing its monopoly status and allows webmasters to disable AIOs for their sites (right now it is tied to featured snippets), I will block AIOs too - not being featured in any AIOs will increase my traffic.

I’m afraid they won’t. The U.S. government’s AI policy is basically a free pass. The FTC is being severely limited and will no longer be able to act. I’m also afraid that this means the antitrust cases are off the table.

[RedBar]

(nobody clicks on the citations anyway)

Known fact or personal assumption?

If I do not recognise the citation source I usually do check it out especially to see if there is anything else of relevance.

[Dooku]

Known fact or personal assumption?

Unfortunately I do not have the website at hand where I read the research, but from memory the CTR from AIO boxes to any website were around 0,2%
So basically useless. Although it was predicted it might go up as more and more search queries produce AIO boxes.
However, seeing as how high the AIO box percentage already is at [advancedwebranking.com...]
How much higher will the CTR still get in the future......I guess negligible.

[Dooku]

This is a really good short documentary that explains why we are in the current situation.
This is REALLY A MUST WATCH documentary. A quote in this video "there are people working at google I would not piss on even if they were on fire!"

Yes, I know some people here still believe in the "goodness" of google, but simply are not aware that everything is by design and scheduled to be "changed" later so that niche, sector, product or whatever, will also ONLY benefit to the profit of google:

Just ignore the Dutch language comments at the beginning....the actual content is in english language:
[youtube.com...]

[Conro]

@Dooku Big companies are like that. I's how that guy who immediately turns his back for his interests. That he is your friend as long as he can get something in return and abandons you as soon as he no longer needs you. Google is exactly that guy, he has shown it several times, but there are those who continue not to want to see the reality of the facts

[RedBar]

This is a really good short documentary that explains why we are in the current situation.

Excellent

[ichthyous]

Whatever was happening earlier this week is definitely over and traffic has reverted. My home page traffic was -25 to -50% until Friday, and since yesterday back to normal. Today traffic is unusually high...

[mosxu]

"en#*$!tification” is probably not enough to describe the #*$!ification that most of us took over the years! The tricks, the personalisation and so on is just peanuts!

[Conro]

On August 2, the AI Act will come into force in Europe and publishers must be able to block bots (Opt-out) that use content from sites to train and create AI content. Does anyone know how Google is moving? At the moment, the ai and indexing bot seems to still be the same and can continue to use the content of the sites for ai overview summaries

[Micha]

@Conro I'd be interested in that too. So far, Google hasn't said anything about it. I wouldn't be surprised if they just try to sit it out for now

[saladtosser]

@Conro how does that affect these AI companies that have already scraped the full internet? Will they need to purge their systems or will this be a, "well whats done is done, moving forward" type approach? In which case 2 little 2 late comes to mind!

[Conro]

@Micha Maybe we will find out this weekend. It is not even known what position Google have taken regarding The code of conduct

@saladtosser As far as I understood then maybe I'm wrong, the "now it's done is done" should only apply to models already produced. For the new models everything is much more complex and obviously in the new models there will also be everything that has been used in the old models or in part. With the ai act, if I have not misunderstood, you must be able to access a public register to know what content has been used and where they were extracted from (e.g. website page) , so that you can request removal If you don't want your information to be used again without authorization. I guess every time you remove info each model created again? For AI companies it would be unsustainable, but even for us it is unsustainable to give away free our work that we have sweated for, so I don't give a damn

[Whitey]

The tools confirm continuing post core update volatility over the weekend:

[seroundtable.com...]

[Dooku]

We regularly see all these research results on well known websites about zero-click, AI boxes, AI mode, CTR from AI......etc....etc...
Why are we not seeing any research results on the number of companies going bust explicitly because of ensh|tification of search results and AI use by google?
One of my websites is service based and many of my customers are in the creative sector.
I have seen quite a lot of their websites go dark and domains not working any more(or already snatched up by brokers).

So, if I am just ONE person that is seeing this, what is the actual scale this is happening now?

[Fluff_Nutz]

The amount of websites getting killed off right now is likely quite high. In, just my sector, I have seen 2 sites go completely dead. One of which closed their site. The other just stopped posting altogether and hasn't been active for several years now.

Other sites I follow are big. Huge in fact but are being hit by tariffs and revenue issues to the point that they did not visit their usual exhibits this past month. I found it incredibly disappointing as they usually do attend those. First time in over a decade. Definitely a wake up sign.

A lot of sites are feeling the pressure. In all niche sectors. Regardless if they sell products or do something else entirely. No one is enjoying any of this, except those who own LLMs and other corporates who are making huge profits right now. Stealing content without a care in the world.

I often feel like just joining them in creating AI content. Why put so much effort into something when it just gets stolen and you end up getting nothing for it. Seems a waste. Alas, I spent a good part of yesterday creating new articles. Up to 30,000+ words worth. Written from my own hand and experience. Why do I do this to myself?

[OldFaces]

I agree with Fluff above. If G continues to rebuild the internet in their image - pushing AI, de-indexing much of the internet, and zero-click search queries - and now with AI browsers such as Perplexity's comet having built in banner advertisement blocking, the Internet Golden Age (where everyone had a 'voice' by publishing websites) is over. I've witnessed a ton of small content publishers like myself who has been doing this for 25 years, simply stop updating their website and change careers and hobbies. This is very sad.

[tangor]

the Internet Golden Age (where everyone had a 'voice' by publishing websites) is over.

This has been true for over a decade. Sad thing is the newbie webmasters are chasing the myth of money and the first generation of webmasters is only now waking up to the fact that myth was never real in the first place.

For all the interest in AI by nation/states the reality is g and other major search engines are actively killing off the small sites (which are generally repetitive/duplicative chasing the same goals in interest of larger/brand sites that pay to play (while getting ripped by AI at the same time). Part of this is to clear out the "noise" of too many sites, expense of "indexing each", and paying penny accounts for a billion when a few thousand REAL DOLLAR accounts are out there.

This is nothing more than a consolidation of a communications niche. Much like what happened after Gutenberg-to-Present---where only a SURVIVING number of publishers (who gobbled up competitors) are 99% and all the rest is fighting for nickles.

THAT SAID, if one is the BEST in the nickles market one can still make a good living by being BEST of the nickles crowd. In another perspective, that "nickle" market is still multi-millions in size because the OTHER 99% is so incredibly large.

ON THE OTHER HAND, g is known to screw things up from time to time. This last update seems to fall into that category and remains unsettled. If it continues the repercussions will resonate to corporate (eventually) that results in a "Fix it Now! The Plebs are in Revolt!"

As an OBSERVATION of g... this rollout did not go well, the collateral damage was unintended and some scrambling is going on behind the scenes. Some will "recover" sooner, others "later", but there will be a change.

Just wonder if corporate's reliance on AI is SLOWING things down because all the scrapped data has not yet been disseminated?

Food for thought.

[christianz]

This is nothing more than a consolidation of a communications niche. Much like what happened after Gutenberg-to-Present---where only a SURVIVING number of publishers (who gobbled up competitors) are 99% and all the rest is fighting for nickles.

THAT SAID, if one is the BEST in the nickles market one can still make a good living by being BEST of the nickles crowd. In another perspective, that "nickle" market is still multi-millions in size because the OTHER 99% is so incredibly large.

Consolidation you describe was HCU. This update and several before it are the opposite of that. July update gave some rankings to small spam/SEO sites (while still reducing total amount of traffic to the web, of course). I have no idea why Google decided it was good idea to weaken spam filtering and undo all progress they made in first half of 2024.

I also don't fully agree with your claim that being best in a small niche can still yield good results. Google SERPs are so irrelevant to keywords that usually its some generic large site or ecom site that will get preference anyway, despite not being relevant and despite being worse in every way.

[haramamba]

I do not see G-bot requests for the last 19 hours on my server.
It feels like the update or bug fixing continues.

[nordland]

A new month is almost over, and I can see record numbers for my website, record highs in search queries and in the number of indexed pages, which makes me feel positive about the future. I also have a record in direct traffic, which I consider one of the most important metrics. I’ve become more active on YouTube, TikTok, and Instagram; even though these platforms don’t drive much direct traffic, they help spread my brand and I’m getting quite a lot of views there. Pinterest also drives some traffic to the site, but nowhere near as much as Google does. I’ve also hit a record in AdSense revenue, which is especially exciting.

[christianz]

For me July will be the worst month in 5 years in terms of ad revenue, down incredible 65% YoY!

[Fluff_Nutz]

@nordland

I hate to break it to you, but Direct Traffic is also bot traffic. Therefore not really reliable. With that said I doubt any traffic source is reliable.

Its all one big risk. A literal online lottery. Two of my biggest traffic sources crashed out this month. Thankfully I'm still +4% over last month in traffic but its incredibly difficult to find sustainable growth.

[haramamba]

For those who like cloudflare.
OpenAI’s ChatGPT Agent casually clicks through “I am not a robot” verification test
[arstechnica.com ]

[Conro]

Google confirms that it will sign the code of conduct for the AI Act in Europe which will take effect in a few days. This means that it will have to provide a way to prevent its bots from using copyrighted content without the author's permission, for example. You will have to provide an opt-out and comply with it. I hope everyone uses the opt out, I have never been against artificial intelligence, but they really exaggerated this time.
[techcrunch.com...]

[tangor]

Consolidation you describe was HCU.

Consolidation continues all the time. Think Highlander (There can be only one.) or Big Fish and Bigger Fish, etc. There's also Diminishing Returns wherein paying out on increasingly smaller "accounts" will be worse than minting a penny. (Currently costs 4¢ to mint the US penny. Currently on chopping block.)

THIS WILL HAPPEN. Not today, maybe not for a few years yet, but it will come. Plan for that. Diversify now or whine later.