April 2026 Google Search Observations

[Martin Ice Web]

from now on I would say we call the updates: core-downgrades – best of two worlds: spam and ads

[edited by: not2easy at 3:07 pm (utc) on Apr 1, 2026]
[edit reason] New month, new thread [/edit]

[RedBar]

but Google took most of the info traffic over a year ago

I feel it was much longer ago than that, certainly 10+ years, this past couple of years have simply been attempts at AI refinement which it has made a right old mess of since, in my industry, they've relied far too much on bad US information!

limiting user's product search results to just ads

Absolutely, Google classified ads have been with us for years now, it's simply a matter of time before organics are consigned to pages 2/3 onwards for popular products.

Has Google actually got the nerve / guts to do this alternatively consign organics to its own tab? That could be interesting!

[Martin Ice Web]

I have the gut feeling that improving a website regarding content, quality and high res pics results in beeing demoted in serps.

Reducing content to a short sentencs + h1 + one pic seems to be promoting the website in serps.

On the other hand, if i look for "Who ist streaming Champions League today?" google shows websites where i have to read and scroll 5 minutes and at the end i still don´t know who is streaming the game?

Shame!

[christianz]

@Martin Ice Web

Have you tried generating AI slop instead of producing valuable content? Google likes slop better.

Did you know that almost all queries in informational sector now have AI slop at position "0", even higher than organic #1? They call it "Gemini".

[Martin Ice Web]

@christianz,

i am complete against writing pointless AI descriptions. I want to share my knowledege about the product. Poeple should see that we have knowledge of what we sell.
We tried AI and we use AI to get hints. But AI is in so many ways bottom line of what realy counts.

And isn´t google telling since years do not write for SE but for customers? Oh, wait. That didn´t work then it it doens´t work now.

Our niche is upside down. All (realy all) my competitors with well informativ descriptions are burried 30+. What ranks #1 to #10 is low content and AI babble.
I though google would like to be a high quality SE.

I miss the days when you used google and you got 5,6 or 8 website to your query. It was very satisfying.
Today I seach with google and get 0 good websites. It results in searching with 3,4,5 attemps with different keywords and phrases. This is not satisfying at all.

[christianz]

I miss the days when you used google and you got 5,6 or 8 website to your query. It was very satisfying.
Today I seach with google and get 0 good websites. It results in searching with 3,4,5 attemps with different keywords and phrases. This is not satisfying at all.

Like I wrote previously - Google is dead. This core update changed absolutely nothing, the previous spam update also changed absolutely nothing. They are done, the slop has won.

Why would anyone under 90 use Google instead of ChatGPT? Back when Google had at least basic ability to detect real websites from fake websites, you would use Google to reach real, organic human content. Now that part is just a mix of AI slop, corporate slop and corporate AI slop. The AI Overview is Temu ChatGPT. Why use Temu ChatGPT when you can use the real one.

[RedBar]

I received an interesting comment from my web host yesterday:

So just some insights:

A lot of people are now finding ecommerce products via CoPilot / ChatGPT / Google Gemini.
They are moving away from the traditional Google/Bing etc

So blocking these scrappers will basically remove your website entirely from being found.
While this is by no means not the “majority” as Google for example deliver results at the top of the page by Google Gemini you will gradually lose more and more share.

This wasn't a generic email, it was a personal one to me which, I feel, shows his lack of knowledge of what is acuallyhappening.

What do you think?

O/T ... Anyone who writes "scrappers" is instantly marked down!

[jellyfish]

Since the core update rolled out, I've seen rollercoaster traffic and a lot of change in ranking stability, even a week after the update has finished. It certainly hasn't settled down for my sites. However long before this update it has been incredibly hard to compete organically, gone are the days when well written and informed content will result in a conversion if you don't 'pay to play'. I haven't banned AI scrapers from my sites, and as a result I now notice AI influencing the source of genuine traffic too.

@RedBar on Cloudflare. You should be able to create up to 5 security rules on the free plan which you can manipulate to block ASNs, Countries, Continents or specific IP addresses too. You also mentioned your mail not sending on some sites. I'm sure you have - but double check your DNS records, to make sure they're all there and if the mail isn't working it might need to be DNS only, not proxied. Hope that helps.

[RedBar]

@jellyfish

Thanks for the CF tips however my problem, therefore theirs too, is that the free service is all mingled-up with subscription stuff. There are simply far too many options thus making it all over-complicated for someone who probably doesn't have the necessary time to learn and understand it, it is over-whelming for the vast majority of users.

To perfomance ... I have three sites with CF now, so far two have seen a good 80% drop in garbage traffic, the other site set-up the same and my biggest traffic site, has only seen about a 10% drop. I find this puzzling since all three sites suffer from the same rogue WordPress probes.

I have no idea why my email stopped with CF and neither has my web host.

I'm going to change web host soon because since the noughties I was with Plesk but for the last couple of years with CPanel and I detest it.

[Micha]

Now that some time has passed, I can say: My news site seems to be one of the winners of the Core Update. The rankings are fluctuating, but Discover has picked up significantly and is now driving good traffic every day. April is already the month with the most readers ever. Though I’m not sure if it’s due to the Core Update or the redesign we implemented after the slump in October, which is now paying off.

The shop is also doing very well; we’re currently seeing steady revenue coming in via Google. Overall, I’m honestly surprised that things are going so well right now.

However, I can also confirm what Martin says: Many of the top results on Google are content-free websites plastered with ads. I don’t understand why Google has pushed them to the top.

[Martin Ice Web]

@Micha, our shop is picking up again too. But at a total different stage. Before the update we sold apples a lot now were are selling pears.

And this update hit ALL of the pages that we reworked to the better. More information, less duplicate content. Straight information. Nice to read. No blabla.
Could be that we triggered a filter : not for normal user but technical versed users. I don´t know.

[christianz]

Ecommerce sites gaining traffic is never a good sign for the organic web in general.

[RedBar]

Many of the top results on Google are content-free websites plastered with ads. I don’t understand why Google has pushed them to the top.

You've answered your own question! Generally clean sites, virtually content free with the relevant keyword(s) and image(s) carrying, usually, Google ads, they're an extension of G basically keeping the visitor in its garden. G does not want independent thoughts, reviews or advice since it cannot monetise such pages but when it does allow such pages to be seen is when someone has a longertailed query since those brief pages do not answer them and, as yet, AI has not been able to digest all that info.

I had a major Cloudflare failure at 02.00 this morning from Ionos, Kansas, 12,258 pages in one go from 66.179.93.249, Range 66.179.92.0/22 added to my list, all attempts thwarted.

[morpheus83]

Since the 14th, Discover and News traffic has plumetted. Anyone else seeing this?

[RubicCubed]

they're an extension of G basically keeping the visitor in its garden.

Talking about walls, I noticed IMDB recently limited access to view movie reviews. To see movie reviews, you must log in now. I bet Amazon, which owns IMDB and limits viewing product reviews in their marketplace, is tired of being ripped off by Google and other AI thieves. While this won't stop all of them, self-defense access barriers are being erected all over the web.

[christianz]

Talking about walls, I noticed IMDB recently limited access to view movie reviews. To see movie reviews, you must log in now. I bet Amazon, which owns IMDB and limits viewing product reviews in their marketplace, is tired of being ripped off by Google and other AI thieves. While this won't stop all of them, self-defense access barriers are being erected all over the web.

I am doing something similar. Increasingly putting more and more pages behind login that were always free to view for everyone. AI thieves are ruining the Web.

[EditorialGuy]

Talking about walls, I noticed IMDB recently limited access to view movie reviews. To see movie reviews, you must log in now.

It's all about gathering and using "first-party data" (such as a site's own subscriber information) for better ad personalization and ad rates. This became a bigger deal when Google was talking about deprecating third-party cookies.

[shawnb61]

This is not a current Google lesson, but one I picked up over the last couple of years I figured I'd share, in case it helps anyone. Possibly old news to many of you.

My site is a forum, like this one. Many threads, some threads have thousands of responses, some have only a couple. My forum software provides each response in a thread with its own link, in case someone wants to share a link to that specific post elsewhere.

In the past, Google would understand posts vs threads. It just ignored the message-level links when crawling. It doesn't ignore them anymore, and this caused a pretty massive increase in the # of requests to our site. Most of which return a 'non-canonical' URL error, because the message level link doesn't match the canonical for that page of that topic. Imagine you have 50 posts on a page; when crawling, googlebot reads that same page 51x and reports non-canonical URL errors 50 times. It only needed to read that page once...

Not all forum software has message-level links, but the ones that do can get hammered by them now. This might also apply to 'report' links - anything per-message on the page.

So, I blocked all message level links in robots.txt. All of those 'non-canonical link' errors became 'blocked by robots.txt' errors in GSC. But... They're not making tens of thousands of unnecessary hits to the site every day anymore. Googlebot traffic returned to normal, < 2K hits/day. Ranking in google search remains very good, in fact, I believe eliminating lots of random errors actually helped ranking.

It's like Google got very stupid over the last couple of years & forgot how to deal with forums. Worse, now they want YOU to parse it out for them. That's the gist of these new "errors" on GSC:

6K+ of your forum pages are missing discussion forum structured data

They used to handle all of this efficiently, without asking us to do their job for them.

[shawnb61]

Lots of talk about CF... I have multiple issues with CF. First & foremost, privacy. They have access to your unencrypted traffic, even when you're on the free plan. To put this another way: They can tell you your customer's passwords. Yours, too. (They published an eye-opening report recently about the current state of password reuse... Eye-opening for all the wrong reasons, as it showed they were analyzing their clients' passwords...) They have many contracts with governments around the world.

And you have deliberately configured them to be the MITM. No need for them to try & sneak there - you put them there.

Second, I have this old-fashioned notion that folks should be able to run their own sites without extra required "security" measures provided by assumed-benevolent corporate entities. Feels like racketeering to me.

And it sets you up in the long term... I predict that, within the next 5 years, CF will either (a) start charging for the "free" plan, or (b) change terms so they can monetize your content in some (seemingly innocuous at first) fashion.

Remember, when you use something for free, you are the product, not the customer.

It's all about trust. You are assuming they're worthy of your trust indefinitely, on into the future. Who knows who owns them... Today & in 5 years...

The thing about slippery slopes is that the first few steps usually appear safe & necessary.

</tinfoil_hat_rant>

[RedBar]

@shawnb61 ... I'm giving CF two more days then removing it. Another site hit by Ionos which I would have "assumed" it should have stopped, fortunately my coding did.

Today's traffic, Sunday, UK has been fine however international has been abysmal at about 25% of average, extremely quiet.

[ichthyous]

It sure is quiet in these forums, I guess everyone has capitulated and realizes that we are now just bobbing along from day to day at the whim of Google's daily changes...it has been stone dead for my business for 8 days now. Only two inquiries from Google and both had such low budgets they got deleted. It is clear as day that all of the viable inquiries now come in in big spurts for about 48 hours right before updates. After that it all dries up...this has been the case for years, but it's getting much worse.

AI sources of traffic: I installed new analytics that track AI Chatbot traffic a little more closely...I am shocked with how many hits per day my site is getting from ChatGPT bot. So far in April 8,178 hits to pages on my site from ChatGPT alone. This resulted in 193 actual visitor sessions, so the clickthrough is 2.36%. That is almost half my CTR for Google web search (currently at a pathetic 5%), but 2.3X Google image search CTR.

All of the Google searches have declined dramatically since AIO were introduced, but they were already on that trajectory for years before. I chose not to block AI bots as it is producing some visits...not nearly as many as I would like considering the number of hits to my site that are required. But it's a new source of traffic, and I will take it. It's only going to grown over time I believe.

Regarding Anthropic / Claude: I signed up for Claude Code at $20/month. I have been completely shocked at how productive and useful it has been for a person like me, who can make edits and knows the basics but who doesn't want to flat out code my Wordpress site templates myself. In the past I had to hire coders for small jobs and it was time consuming, and to be honest they weren't very reliable. This is an infinitely faster and easier experience. The code has been clean and, for the most part, error free. Where there have been errors Claude finds and corrects it. This is going to wipe out freelancers and sites where you can hire coders for smaller jobs completely. There's no way it's not going to impact the profession across the board...fewer programmers will be needed to oversee the use of the AI. Lots and lots of high paying jobs are going to go bye bye, I don't see any way around that.

Regarding Cloudflare: Been using the CF paid plan for years now...at $25 per month if you can't afford that then you really don't have a viable business. It blocks so much of the bot traffic that you (or your client) will probably save money on the reduced bandwidth. On top of that, it speeds up delivery of content quite nicely and has all kinds of tools for optimizing the site which you can take or leave...especially important if you are running a business that depends on sales from outside your own region.

[capulkit]

From last 20 days, Google is deindexing my pages. Have drop of 50% in traffic. I have 10 million pages and not a single of it is spam. Site has been running since 2017 and never had spam issue.

[Micha]

Hmm, I guess I’m the exception. My news site is getting better and better. Discover is driving a lot of traffic, we’re getting more traffic from the news section, and organic traffic is also on the rise. The shop is also running very smoothly right now.

So, I’m just going to enjoy a little peace and quiet for now, until the next chaos hits.

[Martin Ice Web]

Started on friday we saw the biggest drop ever seen. we are right now down to 20%. Replaced by nonsense websites.
google ignores keywords in queries and show complete unrelated results.

@Micha, i am happy for you. Hope it keeps up.

[RedBar]

at $25 per month if you can't afford that then you really don't have a viable business.

For me it's absolutely nothing to do about cost, it's about understanding and being in control of one's websites. If someone wants to outsource and pay for a service that may or may not work, that's fine, however personally I like to solve issues myself which I definitely have done with an excellent piece of code.

I understand that the WordPress issue is not a WP issue per se but a plug-in and / or themes problem.

[Atif]

But honestly, every time there’s a core update, the same pattern happens — sites that were comfortable get reshuffled, and it feels like a downgrade rather than an improvement.

[absheikhh]

I get the frustration, but I don’t think it’s just “Google killing the web.”
What I’m seeing is more of a shift:
Informational traffic is dropping (AI overviews = more zero-click searches)
But intent-based traffic is still there, just harder to capture
Also noticed the same pattern some mentioned, conversions spike right before updates, then drop after rollout. Feels like SERPs get reshuffled and intent gets diluted.
At this point, relying only on Google seems risky. Diversifying traffic (email, social, direct) is starting to feel less optional.

[shadowlight]

I get the frustration, but I don’t think it’s just “Google killing the web.”
What I’m seeing is more of a shift:
Informational traffic is dropping (AI overviews = more zero-click searches)
But intent-based traffic is still there, just harder to capture

AI overviews = more-zero click searches because google is re-writing and displaying web site / business created content on their own platform which ultimately results in less traffic / opportunities for the creators of the content to capture value from, and more opportunity for google to capture value from said content.

Informational and commercial intent based traffic are both still there, the shift is exactly what you describe. It is harder to capture, because google have made it harder to capture.

They have done this by displaying third party content on their own platform (AIO), increasing zero click searches and cluttering the SERPS with PAA, PASF, Find Products From Advertisers, alongside ADs.

This results in information that requires more effort to find, increasing both keyword search volume and AD impressions / clicks.

The SERPS, especially on mobile are extremely AD heavy, especially for anything commercial. Often 7 ads before you reach the majority of organic listings on mobile search and that's in addition to AIO, PASF, PAA, Find Products From Advertisers with one or two organics mixed in and two of these organic results are not even as big as ONE single Ad.

Depending on search term, their affiliate SERP features might also be present (Flights, Hotels, Shopping).

This leads to many businesses either laying staff off, shutting shop, spending heavily on ADS often with diminishing ROI and passing the increased cost to consumers.

At any point relying on google for traffic is extremely risky, unless you have the budget to buy a steady stream of traffic from them and unfortunately they control 90% or so of online information discovery worldwide.

[RedBar]

The SERPS, especially on mobile are extremely AD heavy, especially for anything commercial.

Absolutely plus usually very localised. I try not to business search on mobile, it is far too awkward.

[shawnb61]

Who knows, I may turn to CF one day. Thus far, I've been resisting.

I'm surprised the dangers & tradeoffs of CF are just not discussed at all, anywhere. Folks seem unaware of the whole MITM aspect, and the potential issues today & down the road.

I've been keeping a very simple approach: I block ASNs with bad guys in them - ***unless*** I have active registered users there.

My experience is that ASNs exhibit bad behavior - at the ASN level. Bad activity comes from them - aggressive crawls, hack attempts. Either they are small regional orgs who are overwhelmed (lots of that in Brazil last year, I suspect), or, they simply don't care what their customers are doing as long as they pay their bills. One way or another, bad activity comes from that ASN. Blocking part of an ASN doesn't help. That activity just moves to other IPs within the ASN at a later date.

It's the org that's the problem, not the IP. Whether it's lack of governance or resources, the result is the same.

To complicate matters, the IP ranges within ASNs change over time. A LOT... ESPECIALLY for these bad guys... I've blocked whole ASNS, only to find myself under attack by them a few weeks later. "I thought I'd blocked that guy..."

But I had blocked the whole ASN via a bunch of CIDRs, as it was defined on a particular date. But that definition changed a couple weeks later.

Before my ISP allowed blocking by ASN, I blocked by CIDRs for my problematic list of ASNs, and found that I had to re-cast my set of CIDR blocks ***every month*** to follow the new ASN definitions. Things improved quite a bit, and I didn't have to chase specific IP ranges anymore.

My dilemma occurs when I have active, registered, participating users accessing my site from a bad ASN.

Which brings me back to CF... I have lots of active, valid, users in various bad ASNs. Contabo, Datacamp, Digital Ocean, AT&T, AWS, AS45899 (Vietnam), etc. In my experience, blocking by IP range is a waste of time, and at the same time I don't want to block those ASNs due to the active, participating, valuable users I have there.

So... I may ultimately need to fall back to CF some time. I haven't given up the battle just yet...

[shawnb61]

Not everybody is running an ecommerce site... My site is just a bunch of guitar junkies sharing info on guitar effects, guitar synthesizers & modelers. 44K members, globally. Zero monetization, a couple of us just renew once a year & ask folks to help offset & they do.

There are many, many thousands of such communities out there like this, running their own sites, having somehow avoided being Borg'd into the social media platforms. My site uses SMF, and I help out over on SMF, supporting these sites. "Labors of love" - train junkies, gamers, model junkies, fantasy ethusiasts, anime fans, photographers, developer communities, small local manufacturers, artists, neighborhood orgs, churches, etc. Most of these little sites have hundreds of users & low/zero monetization.

Many created their sites 20 years ago and have been keeping them afloat. Fighting ongoing changes to PHP, MySQL, Apache, etc., over the years is bad enough. Now folks need a much higher level of technical skills to keep such sites afloat as they are under siege by bots. Or they need to shell out more money, as the crawling blows past the CPU or other resource limits for the cheapie hosting plans we all use.

I am convinced in the long run that this federated model is going to regain relevance, as folks realize they need to abandon the social media platforms. Reclaim their content.

But in the meanwhile, IMO, we gotta help these very cool islands of enthusiasts stay afloat. Whether they're commercially viable or not.