January 2024 Google Search Observations

[Micha]

---

System: The following message was cut out of thread at: https://www.webmasterworld.com/google/5098123.htm [webmasterworld.com] by engine - 9:24 am on Jan 2, 2024 (utc 0)

---

Happy New Year

I hope you survived New Year's Eve well and that your websites did too.

Apparently the upswing continues, I continue to see a slight improvement in the ranking and the number of readers. (on average +11.2 percent more per day at the moment).

It may be slow, but I hope it continues.

[EditorialGuy]

The popular opinion is that they want to put all their efforts into AI (in which they are still an also ran). I wonder if this will prove to be another 'other bet' eventually?

"Artificial intelligence" covers a lot of territory. Will Google rely increasingly on AI as an internal tool for creating search results? I'd say that's a given. Will Google bet the farm on AI as a core end-user product? That's less clear (and maybe less likely).

[superclown2]

Will Google rely increasingly on AI as an internal tool for creating search results? I'd say that's a given

Their current system produces profits beyond the dreams of avarice. Why risk it for switching to another system that costs enormous sums to run? And aren't they facing enough lawsuits already without their search results containing catastrophic hallucinations?

In the meanwhile they are still shedding staff and slashing costs on perks that employees used to take for granted. AI is expected to replace many humans in the ad sales departments but will that really be an improvement? It's difficult enough already to contact a real human there.

Yes I can see them using AI to reduce headcount still further but what effect that will have on employee morale and efficiency remains to be seen. In the meanwhile I still don't know anyone who uses Bard, except when ChatGPT is down. Like Waymo, it might produce profits one day, or may go the way of Google Glass. I expect more redundancies will follow as they ditch more loss making 'other bets'.

Frankly I think that this obsession with AI is a huge gamble that could blow up in their faces. It only takes a few court decisions to side with content creators who produced the data that these systems were trained on to derail the whole project - and governments, who don't understand it, fear it so legislators are unlikely to be friendly toward it. Plus nearly everything they touch which isn't ad related seems to turn to dust.

Finally: Bing launched AI based search to great fanfare. After an initial small rise their share of the search market seems to have fallen rather than risen. Do the public really want it? Gen Z don't seem to, they prefer to get their info from Reddit or TikTok; and we older generation still see it as a useful tool for many things, but looking for the cheapest widgets? No thanks. We prefer to do our own research.

[RubicCubed]

We prefer to do our own research.

While I agree, the unfortunate problem is these search engines no longer care about what their users want. Search engines will continue for force feed their users whatever makes them the most money with little regard to the search experience and quality of the results presented to them.

I find all the ads in Google, their refinement boxes and of course Google's extremely poor organic results to make search extraordinarily frustrating so I don't use Google. When I go to Bing, and accidentally click on their PAA box and get an "Ask Bing Ai" response, I quickly back out of it and then click on the external link below it. The time has come for a paid search subscription service so we can choose our own layouts free of ads, videos, refinement boxes, etc. Also in a search subscription service we could be given more control to blacklist domains from the results we receive (amazon, reddit, outlookindia, etc.). But empowering users, by giving each user choice and control of their search experience, is not something any major search engine is apparently willing to do. Why won't any major search engine offer a subscription? My guess is one accidental click on an ad, depending on the search, can easily fetch the search engine $10+ in revenue. It's possible the "average" user clicks on ads valued at $50+ each month and few would be willing to pay that much or more for a subscription to get rid of ads and all the garbage that has littered the search results.

[ichthyous]

Massive drop today...search traffic is -51% at 1pm. USA traffic -61%, UK is -68%! Looks like something is rolling out...

[EditorialGuy]

Yes I can see them using AI to reduce headcount still further but what effect that will have on employee morale and efficiency remains to be seen.

I don't think the point of using AI in Google Search is to reduce headcount, it's to improve search results by slicing and dicing data to a degree that wasn't possible until machine learning came along.

The effectiveness of using AI to improve search results is mixed (at best) so far, but I'd imagine that Google is willing to tolerate crap in the short run for success in the long run. We're now living in an era when beta testing is ongoing and done in the wild. (That's the price of "free.")

[ichthyous]

Is anyone noticing all of the spam in Google news feed? There are now two fake news articles that come up at the top when I google my own name that simply redirect to ecommerce sites. As if anyone would buy anything from an ecomm site (selling shoes) using those kinds of spammy tactics.

[superclown2]

The time has come for a paid search subscription service so we can choose our own layouts free of ads, videos, refinement boxes, etc.

There was one called Neeva and it was very good. According to the owners it failed not because people wouldn't pay but because no-one could find them thanks to Google's monopolistic activities.

I don't think the point of using AI in Google Search is to reduce headcount, it's to improve search results by slicing and dicing data to a degree that wasn't possible until machine learning came along.

They were quite capable of creating superb search results 20 years ago without the expense of massive data centres and enough electricity to power a small country. That was before Wall Street made them put profits above quality of course. Would they really want to improve their results at the cost of fewer clicks on ads?

[EditorialGuy]

They were quite capable of creating superb search results 20 years ago without the expense of massive data centres and enough electricity to power a small country.

The Web was a lot smaller (and a lot simpler) 20 years ago.

That was before Wall Street made them put profits above quality of course. Would they really want to improve their results at the cost of fewer clicks on ads?

That would depend on how much revenue the ads were generating. Overloading pages with ads can yield diminishing returns.

[Micha]

@Ichthyos Many people have already noticed the spam, and the problem is getting worse. Plus: there are fewer and fewer really new articles in the feed.

The decline continues for me too. Almost 1,000 keywords have been lost this week. I find it particularly annoying that some of the top-ranked websites have been replaced by off-topic websites that have nothing to do with the search term.

[RedBar]

I'm having to remove Statcounter because today 42.2% of my PVs are supposedly from Huawei Clouds Singapore and no matter what I do I can not seem to be able to block their search bots.

The bizarre thing is my Awstats, Webalizer and a couple of other analysis programmes are not showing anything from Singapore whatsoever.

[RedBar]

Fingers crossed I've just found some filters in Statcounter, we'll see if this works.

[ichthyous]

The decline continues for me too. Almost 1,000 keywords have been lost this week. I find it particularly annoying that some of the top-ranked websites have been replaced by off-topic websites that have nothing to do with the search term.

I've had a steady loss in top ranking terms since early November. That followed a steady climb for months starting in March 2023. It strikes me that this is some sort of cycling as the terms that are being shed are the ones gained over the previous six months. In my case it coincides with losing a lot of high powered inbound links. Perhaps you can check to see if you have also lost some important links from high DA sites. It's a lot of work but replacing them helps to turn the trend around.

[Micha]

@Ichthyos I had a look, my links are okay, a few high-quality backlinks have gone, but some have also been added. I assume it's the usual Google nonsense that G doesn't even understand himself.

[Dooku]

@Redbar, Would you mind explaining why you are filtering in Statcounter if your aim is to block that bot altogether?
I mean Statcounter filters just remove that entry from your stats.....while that bot is still hitting your website.
I might be completely wrong....still, so just asking.

[NeapTide]

After few days of getting good traffic the same PAA boxes issue has started again pushing down serps. 1 organic result followed by PAA then 1 or 2 results followed by big PAA box then 2 or 3 results and another PAA box.

Google is now search engine for just 1 or 2 organic results. Rest can just go to hell (google's new motive)

[RubicCubed]

1 organic result followed by PAA then 1 or 2 results followed by big PAA box then 2 or 3 results and another PAA box.

The PAA box now appears above the first organic result for many of our keywords, though its impact on traffic is negligible since the four ads above it are siphoning off most of the traffic. Technically SEO isn't dead, but the traffic rewards for being ranked #1 are so insignificant only those who use AI/automation to generate mass quantities of pages will realize any profit from it.

[superclown2]

Just as I thought results couldn't get worse Google proves me wrong again. It's as though they are playing bingo and picking results out of a hat.

Just one example; a search for 'best red widgets' (where red widgets is a popular search term). The organic results are:

1) Review site 1
2) Review site 2
3) A site that is faintly related to the search term but which doesn't say anything about it
4) Review site 2 again
5) Review site 3 (a German site with over 8 million super spammy backlinks; I'm in London)
6) Review site 3 again
7) A major newspaper with an article about a slightly related but otherwise different subject
8) Reddit
9) A site that sells widgets but not red ones
10) A site providing customer service addresses and phone numbers

Page two was almost as bad with not a single site about the search term I entered. I gave up at this stage.

I stress that the search term I entered was 'Best (Very Popular Search Term)'; I didn't add any more words and I'd deleted cookies beforehand.

No doubt they are in panic mode to hit their earnings target this quarter.

[RedBar]

@Dooku

@Redbar, Would you mind explaining why you are filtering in Statcounter if your aim is to block that bot altogether?

Quite simply because it was significatly skewing those stats.

I tried for a couple of days filtering through Statcounter, nothing whatsoever worked therefore Statcounter has been removed.

I have tried for some 7-10 days now through htaccess and robots.txt, nothing whatsoever has worked, even blocking the Singaporean IP completely has failed, nothing is taking any notice of what I am doing.

I have read more suggestions and ideas about what to do that it is simply becoming ridiculous, some of these companies are ignoring all "supposed" rules and agreements.

To say I am pi$$ed off with the entire culture of all this cr@p at present would be an understatement.

Plus you will be seeing a lot, lot less of me here simply because no matter what any one of us, or collectively, says or does, no one is listening, no one will do anything, we're all just pi$$ing into the wind trying to vent our frustrations as we sink into their abyss.

[ichthyous]

I am seeing stronger traffic in general, despite more bleeding of top 3/10 ranking terms. It had stopped for a week, but now the loss of terms has started again. One thing I see is that non-USA traffic has pretty much returned to normal or surpassing previous highs, while USA traffic remains stuck at much lower levels since early November and has never recovered.

Upon further inspection I see that is mostly due to an 80% drop in traffic from a specific set of terms that was brining in a lot of daily visits. However, my ranking has not changed at all for these terms. I did make some sitewide changes during this time, increasing the number of related posts on the page from 6 to 10...which caused the pages to drop from green to yellow in Core Web Vitals, but it's also getting visitors looking at more content.

And I updated every single page with new titles and sometimes descriptions to conform with limits on the length. So often shorter and more specific titles. That took two months of work as I did each one by hand to ensure that it wasn't cookie cutter, and made sure the related posts were truly semantically related.

[Micha]

Oh well, today is another one of those days when you wonder whether Google still lists your website at all. G's traffic almost came to a standstill from yesterday to today, and I'm still losing keywords.

So it's business as usual, it goes up for a few days, then it goes down again and so on. Overall, however, the number of visitors remains at a constant level, albeit significantly lower than in January 2023.

[superclown2]

no one is listening, no one will do anything, we're all just pi$$ing into the wind trying to vent our frustrations as we sink into their abyss.

I wouldn't get too downhearted. This year the EU Digital Marketing Act and the UK Digital Markets, Competition and Consumers Bill come into full force and they both have real sharp teeth. OK the USA litigation will probably drag on for years but the rest of the world (including, now, Japan) are not as patient or prone to 'lobbying'.

I am now finding ChatGPT is failing to answer a lot of queries that they were perfectly happy with a short while ago. I'm wondering if this is as a result of all the copyright issues that are going through the courts. Perhaps AI will prove to be a damp squib, as far as search is concerned, after all.

In the meanwhile I am seeing huge swings in the Google SERPs from day to day with many of my search terms oscillating, seemingly, between the top of page one to buried in page 3. Big difference between mobile and desktop too. However this doesn't seem to make the slightest difference to click-through rate so I am wondering if the influence of localised searching has been ratcheted up. My mobile IP address, along with it's apparent location, changes whenever I turn the phone off then back on again (and I constantly delete cookies); one minute Google says I'm in London then the next I'm in Glasgow.

[Micha]

Meanwhile, a study from Germany proves that Google search really has gotten worse. In other words, it confirms what we have often said. Link to the study: https://downloads.webis.de/publications/papers/bevendorff_2024a.pdf

[edited by: not2easy at 2:28 pm (utc) on Jan 17, 2024]
[edit reason] delinked for pdf clarity [/edit]

[christianz]

Oh well, today is another one of those days when you wonder whether Google still lists your website at all. G's traffic almost came to a standstill from yesterday to today, and I'm still losing keywords.

Can confirm. Yesterday and today so far are really bad. Just like last week.

[Dooku]

@Redbar..... I know what you mean :-(
But still want to provide some help, also for other members coping with the same problems.
IP addresses of some bots or malicious actors often are not blocked by htaccess or even the linux firewall on your server.
In that case ad the IP address to the Nginx blocklist if you are using an Nginx based webserver......often this will work.

But over the course of several months I have added almost all my websites to Cloudflare and have created extensive WAV security rules in it's firewall.
The free version of cloudflare is enough for this for most users. It's a lot of work at first yes, but from then on it needs minor up keep when you notice a new bad bot or malicious actor.

I suggest blocking all countries that are just not relevant first.....from there go down in more detail from ASN network numbers, ip address ranges, user agents etc...etc...
You can see easily from the server's raw logs if anything is malicious or not or is just useless. If in doubt copy and paste its IP address in abuseipdb.com and you will quickly see if you need to block it or not. For more detail use any Whois tool you like and from their any IP to ASN converter tool for their network range.

I have noticed a major decrease in load on the servers because of this as the amount of bad bots, scrapers and especially script kiddies and hackers have increased exponentially over the last 10 years......thanks to youtube "tutorials".
But according to my stats it also improves the bounce rates by removing all those "1 to 3 second one page" visits from useless non-human visitors. And because of that, but this I can not prove, I have the impression many pages got ranking improvements as the overall visit length increased and maybe google thinks the page is now more relevant?(without changing any content).

I am just providing this info from my own experience, maybe it will help some of you. We need every little bit to stay afloat in the current SE landscape.

[christianz]

Lower bounce rate and higher average session time can potentially mean your content is worse not better than some other content with bouncy traffic and short sessions...

Maybe it is filled with ads, popups, confusing navigation and content that you have to click at least once (in hope) to see what you were actually looking for versus straight-to-the-point page that immediately provides what you want and, therefore, generates mostly single page view (bouncy) traffic...

[EditorialGuy]

Lower bounce rate and higher average session time can potentially mean your content is worse not better than some other content with bouncy traffic and short sessions...

As with most things, it depends--not only on the content, but also on the visitor's intentions or interests. That's why Google has said for years that bounce rate is a "noisy" metric.

To put it another way, how is Google supposed to know whether Bubba spent five minutes on a page about Widgetville pedicab fares because (a) he had trouble finding the information, (b) he was interested in the details [fares for kids, fares for disabled people, day passes, etc.], or (c) he stopped to brew a cup of coffee before looking at another search result?

[ichthyous]

Yep another big drop in traffic...you could see it kick in yesterday eve as European traffic dropped first, then today USA traffic is -33% from an average Wednesday...and it has stayed at exactly -33% since 11am, so seven hours in a row. That is statistically impossible, so it's all just complete manipulation. It looks like a new pattern of one day on, one day off is becoming the norm.

[KaseyM]

Dropping ever since October and no sign of slowing down.

I've never seen the results as bad as they are now. A complete mess and some Indian newspaper is just ripping my content and ranking about me within hours of publishing.

Helpless.

[ichthyous]

@KaseyM Perhaps block all traffic from India and they won't be able to reach your site to rip it off? May not work entirely but will slow them down. I have blocked China entirely due to concerns that their AI bots will just completely rip my content. About 6 other countries totally blocked as well, and the list is growing. I don't block India, but I do have them on "managed challenge" via cloudflare. A lot of Indian traffic is just scouting for contact forms etc

[christianz]

Helpless.

There is nothing we can do and nothing we should do. It's up to Google to get their act together.