Forgot your password?
typodupeerror
Links AI Google The Internet

Publishers Are Losing Google Traffic As AI Answers Replace Links (axios.com) 83

alternative_right shares a report from Axios: Google has basically stopped sending people to websites (including our site) for answers and information. Instead, it's using AI to answer them on its platform, in its words. Chartbeat data shared with Axios shows Google Search traffic to publishers fell 34% over the past year. That pain is regressive. Over the past two years, small publishers lost 60% of referrals from search overall, medium publishers 47%, large publishers 22%.

Publishers Are Losing Google Traffic As AI Answers Replace Links

Comments Filter:
  • Ads and Bad Advice (Score:5, Insightful)

    by darkain ( 749283 ) on Friday July 31, 2026 @02:08PM (#66266516) Homepage

    1) nobody wants to read an "article" that has more ads than content when researching a topic (yes, ablock exists for us techies and I have it, but talking in the more broad common person sense).

    2) nobody wants to read a forum thread with over 100 posts to try to find the one post with the real answer surrounded by a collection of bad answers and people saying "YUP, IT WORKS NOW" with zero context otherwise.

    Make your content on-topic and valuable, and users will return. it is as simple as that. These are just two of the main things I see as devauluation of a lot of these web sites complaining about a drop in traffic. They're most certainly not the ONLY reasons, but it is two of the major points!

    • Bollocks. Just spend more spamming your products or links everywhere. Or pay someone to guess whatever the current ranking system is to game it. Doesn't replace what is happening to traffic. AI is in almost every "people also ask" after the main search, a good portion of the time it still contradicts the links it uses as citations. Quality of site has little to do with it or sites like quora would be nowhere near the top.
    • by taustin ( 171655 )

      1) nobody wants to read an "article" that has more ads than content when researching a topic

      And often, those ads cover the entire screen, and are four or five deep. If it - literally - takes me longer to close all the ads than to read the article, you deserve to go out of business.

      • Those who put information on the internet do so on their own dime. They must pay for the hosting, not to mention whatever it took to come up with the content. They can't do that for free.

        The ad model is a bad one and it is right in the middle of a continuing arms race. People hate the ads so they use ad blockers, and sites try to use ad-blocker-blockers, and people try to mentally train themselves to ignore the ads, and sites throw obnoxious ads right on top of the content you are reading (not to mention

        • by taustin ( 171655 )

          What's your point? Since most web sites are completely useless now, the only difference between that and them going out of business is that I waste less time going to completely useless web sites.

    • Make your content on-topic and valuable

      Content is king. I tend to agree with you. However, I think anyone can look clear eyed the situation and see that "users" are going to do pretty much exactly as you say shortcut the content-slog with an AI answer. So, I'm still thinking web models for advertising are going to struggle as well as the web, in general.

      Here's where I go off the reservation though. See, I loved the internet in the late 1990's. Everyone was going to get their own IP space. It was going to be theirs (because hey, we have 4B addr

      • Lately, I've been finding myself deleting the internet from my life. I think AI has some amazing uses, but the thing I really need less of is the strip mall. Google searches have completely fallen apart. Google's AI answers are still hallucinations most of the time. The one thing Google still does well is the thing I wish it would stop doing, which is indexing people's private information. Other search engines are far more respectful of people's privacy.
    • by Voyager529 ( 1363959 ) <voyager529 AT yahoo DOT com> on Friday July 31, 2026 @03:03PM (#66266650)

      nobody wants to read an "article" that has more ads than content when researching a topic

      This. This right here is the answer.

      Google search originally was creepy-accurate because it was built on the work that was done elsewhere. While other search engines looked at frequency of search terms and other easy-to-game metrics, Google did it based on links, meaning more linked-to pages got preferred rankings. That worked for a very long time, until that too got automated and SEO slop took over search indices. Google itself began to have more and more ads creep into search results, and reward clearly-seo-slop over the sort of sites people actually wanted to find. Sites got so plastered in ads and tracking scripts that they became utterly unusable without an ad blocker, even for people who WANTED to support the sites...then there were the cookie banners, then there were the modal overlays...and if you had a labor-of-love website that legitimately had useful, human-created or curated content that was useful and functional...you were buried on page six of the search results because you were competing against the SEO slop and backlink farms.

      and then ChatGPT/Claude/Gemini/Perplexity came around and were able to sift through all the crap to give everyone the paragraph they were actually looking for in the first place.

      You dug your own grave, Axios...when your website takes a minute to load its moving video ads, people tolerate it in the absence of an alternative. Gemini is terrible for 101 other reasons...but if it puts the slop sites out of business, I'll accept the tradeoff.

      • by sinij ( 911942 )
        You are mistaken in your analysis. It is only better for now, until enshitification lands you right back into ad hell, only it is much harder to block because it is all single-source with content you want, unlike ad networks that are separate and identifiable.
      • Axios dug its grave by shoveling out totally fabricated news to help insiders game the stock market.
      • Sites got so plastered in ads and tracking scripts that they became utterly unusable without an ad blocker, even for people who WANTED to support the sites

        Google is responsible here. The could penalize websites that have too many ads. But they won't because that is how they make money.

        It's a clear example of a monopoly making the world worse.

      • The Internet was built on ads, itâ(TM)s what paid for sites for many years. If the content publishers had stuck with decently built ads that scrolled with the page and didnâ(TM)t set your cpu on fire maybe ad blockers wouldnâ(TM)t have become so popular and we wouldnâ(TM)t be where we are now.
    • What's the alternative? Glue on pizza? https://dailydot.com/google-se... [dailydot.com]
    • 2) nobody wants to read a forum thread with over 100 posts to try to find the one post with the real answer surrounded by a collection of bad answers and people saying "YUP, IT WORKS NOW" with zero context otherwise.

      This has been a solved problem for a long time with quite often the correct answer being highlighted or original posts updated. But that's beside the point. If you trust the AI answer, you're not going to click an article or a forum link regardless of whether they show ads or highlight a solution.

      Though sometimes you do need to dig deeper. Google's AI never answered WHICH glue to use to keep cheese from sliding off a pizza, you'll need to actually read the reddit post for that.

    • Honestly, Google's search results are horrid as well. Search any news story, the first 100 links will be the EXACT same text mixed with 100 different forms of advertising bulletin boards.
    • by jmke ( 776334 )
      >Make your content on-topic and valuable, and users will return.

      nope they won't. have zero ads on my site of 20+ years, plenty of content, but AI is just delivering the summaries on their platforms, nobody visits the source. AI is what google dreamt off when it launched their first google search frontpage, and have been working towards all this time, keeping people on google sites and not letting users visit other sites. GG: they won.
  • because you don't have any idea where it got its "facts" from.

    With a link you can (roughly) judge the credibility of the source. I.e., did the "facts" come from published research or some nobody on social media. Or some known propaganda outlet.

    • GIGO: Garbage In, Garbage Out. As AI increasingly becomes trained on AI output, it will be increasingly less connected to objective reality.
      • GIGO: Garbage In, Garbage Out. As AI increasingly becomes trained on AI output, it will be increasingly less connected to objective reality.

        Speaking of garbage in, am I confused or did Google just basically replace their for-profit advertiser revenue streams with..themselves? How exactly does Google make money using Google AI here?

        The garbage out(put), might soon be the least of their concerns..

        • Won't be long before more obvious ads start appearing in the AI overview. They can continue their revenue without hugely increasing the advertising on page now. Which is ideal for them as now all ad revenue is theirs. I can't see anything that could go wrong with this.
        • Haven't you heard? The AI business model is that now everybody makes a living doing each other's laundry.
    • My experience with AI is that when it can't find the answer, it just makes one up instead of admitting it doesn't know the answer. Unfortunately, I have worked with people like that too. It is not productive.
    • by allo ( 1728082 )

      Google puts favicons of sites next to its answer, perplexity also links all the sources. If you need to know more, click the source links.
      Google and duckduckgo just stupidly summarize the top search results (thus also amplifying bullshit in the results), perplexity follows more links than just the first level from the search results, but also adds citations to its claims.

      • by alcmena ( 312085 )

        > If you need to know more, click the source links.

        Do you actually do that often? I have, and many times the information in the summarized answer doesn't exist in any of the cited sources. I strongly suspect it's because the AI doesn't "know" where much of the answer truly came from, or even if it's real. It just "assumes" the top links are probably the source.

        • by allo ( 1728082 )

          Quite a few times, as the summaries generated are mostly just that, summaries. For more information you either read the related links or ask a full chatbot. People always say "Gemini sucks" and think of the box above their search results, but that's literally the result of their smallest model.

    • by Bahbus ( 1180627 )

      This is just patently false.

    • by gweihir ( 88907 )

      Exactly. And AI "Answers" keep telling me the most ridiculous crap. I think it must be something like > 30% hallucination for the things I search.

      Example: It gave me melting points for a plastic at 100C and 350C, both (!) in the same posting. The real value for that plastic is around 215C. Completely unusable.

  • Yesterday I got an unwanted Gemini result at work where I can't block them that wrote the word "covered" as "covered". That was fast from the only error as well. Public Gemini is clearly the most shit of all LLMs.

  • by TheMiddleRoad ( 1153113 ) on Friday July 31, 2026 @02:20PM (#66266550)
    Search quality has dropped off so far over the past several years that it's almost useless. Too much AI slop out there, apparently. Gemini gives useful answers. I set up an agent that queries gemini and serves those results immediately then looks at what gemini says and does its own deeper search. There are usually significant differences and improvements with the deeper search. Or I can spend 20 minutes digging through forum posts and spam websites, trying to find the answers that might be eight links down or not there at all. It's a fucking mess.
    • Or I can spend 20 minutes digging through forum posts and spam websites, trying to find the answers that might be eight links down or not there at all.

      ...because they've configured the web server to let Google index everything but it's a paywall for humans.

    • The only issue , though, is as we go forward. The quality of results will fall off. If a website or forum can't get anything but AI scouring it, they'll simply cease to exist, and AI output ( based on input) will fall off considerably.

  • ...but we got there, Google is finally doing what ask.com was doing in the 2000's.

  • Everyone's acting like Google stole something uniquely valuable.

    The uncomfortable reality is that for most searches, people don't need your version of the answer. They need an answer. And that answer has already been written thousands of times by thousands of websites.

    LLMs compress those thousands of near-identical explanations into one response. That's not a bug -- it's exactly what they're good at.

    The SEO era rewarded being the 500th person to explain "What is X?" The AI era won't.

    If your business

    • This is exactly what someone would say if they never produced any content of value

      • by allo ( 1728082 )

        Do you think the 101st answer for a question answered 100 times before is of value?
        I think the parent phrased it right. You get now rewarded for contributing new knowledge, not for parroting something said a hundred times before. Providing something new unfortunately means you need to do more than just find one of the 100 answers, change the writing style and put your own sponsor links in the text to score online points. Maybe we get back to the old web, where people published online because they wanted to

      • This is exactly what someone would say if they never produced any content of value

        The fact that this is unintentionally self-referential makes it even more funny.

      • Recently, the search traffic to my website has decreased by roughly 10 to 20 times, and my well-researched original articles, which have not been published anywhere else, range from three to ten full pages, so Gemini cannot even summarize them properly. Some articles have been translated into multiple languages.

        In the meantime Google removed me from the AdSense publisher program a couple of years ago, as they did with most Russian publishers. And they refuse to accept me back, even with a different account

    • except: why produce anything of value if AI is simply going to suck it up and reserve it to people? Why spend time on "original reporting, proprietary data, real expertise, or genuinely distinctive analysis" if in the end, everyone will ask AI and AI will repackage your work?

      • except: why produce anything of value if AI is simply going to suck it up and reserve it to people?

        If what you create/publish can be presented as a three-line AI summary, it may not be worthy content.

  • by supabeast! ( 84658 ) on Friday July 31, 2026 @02:35PM (#66266586)

    If I need to know something and I search with Google it kind of sucks. Half of the articles are clickbait, written by non-experts or a shitty AI that was given a garbage prompt. When I go to the web site I immediately have to jump through as many as three popups about cookies. Then, ten seconds after I start reading, a full page popup appears compelling me to sign up for a mailing list. If I am on a mobile browser with an ad blocker the page may refuse to load. On desktop I might get a popup telling me to disable my ad blocker. After all that shit it turns out that the garbage article does not even answer my question. So I have to keep doing this with every link until I either find an answer or give up.

    But if I ask Claude it will usually give me a great answer. Sometimes it gives an entire essay. The writing is much better than the garbage I would get from a Google search.

    But lately I have taken to just looking it up in books. This month I bought a big Merriam Webster dictionary along with the Oxford dictionaries of economics, finance and banking, and politics and international relations. It's great! Entries are short because Oxford is not cooking up five+ paragraph articles to put ads into. The writers were experts. And they were human. I still resort to Claude at times, and it can be faster, but I like ink on paper.

    • by 0123456 ( 636235 )

      Except the answer is often wrong.

      For example, last year I was looking for info about a library that some old software we maintain every decade or so uses. The "AI" told me it was deprecated, but when I actually went through the search results it turned out there was some other library of a similar name which was now deprecated and end-of-life.

      Without the search results you can't check to see whether what it told you us is true or misinterpretation or just plain made up.

      • by allo ( 1728082 )

        What is "the AI"? You need to specify the model (and version or at least date so one can estimate what version it was) for a useful critique.

      • by Bahbus ( 1180627 )

        Except the answer is often wrong.

        You need more recent examples for how fast AI models improve. Try that search again and I bet you get a different result.

        • by 0123456 ( 636235 )

          Except they still do it. That was just a glaring example that sprang to mind.

          Another example more recently is going around in circles trying to find out how to change some setting because the "AI" (Google, Bing, whatever) pulls instructions from two years ago and the UI has completely changed since then. Or maybe it just made up UI features that never existed.

          Oh yeah, and then there was Copilot recently, I asked it to suggest an app which would do something, it told me there were two free Microsoft apps whi

          • by Bahbus ( 1180627 )

            The first is a problem of information discovery/association, assuming that the correct information is out there. The second is a problem of Copilot not really being good at that type of thing, depending on the feature needed. And both can be exasperated by user incompetence (not saying you are, just in general).

            Google AI is correct, for me, about 90% of the time. 9 out of the other 10% of times, it's simply because I forgot to be specific enough (generic Linux results instead of distro specific, for example

          • by allo ( 1728082 )

            What's so bad about RTFM to find a setting? AI is good for finding solutions and good for well-established knowledge (at least larger models). Things like where exactly an UI setting is, is neither. The problem starts with how often this is in the training data. You hopefully find it in TFM. Mostly that's it. How many sites do you know, that tell you where to find that setting, except the manual? Where should the knowledge come from, when it is that rare on the web and possibly not even in the training set?

    • I bought a big Merriam Webster dictionary along with...

      You may be an exception. The books you mentioned are reference books. I am worried that reference books will disappear, because most people won't use them. Which means that we will be at the mercy of getting our information from AI.

      Reference book publishers might publish their books through an LLM: you can ask questions but nobody gets the whole book.

  • Follow a search link and you wind up on a page that has an opening summary paragraph sort of about the topic. Then goes on for 6 more paragraphs (with an image or two between each paragraph) that says the exact same thing as the summary using different wording while never actually offering any details about what brought you to the stupid page.

    Slop ruined search links before AI, hard to care about someone not getting search hits anymore. I guess publishing is not a profitable business anymore - oh well. M
  • by allo ( 1728082 ) on Friday July 31, 2026 @02:44PM (#66266614)

    If your article can be replaced with a three line summary, your article is not worth my click.

    • Now the real question: If the article didn't exist, where will Google get it's summary from?

      I get your complaint, but you should really consider the end game here.

      • by allo ( 1728082 )

        Hopefully from a better site. Even though it may contain a bit of information buried under a lot of fluff and ads, the information in clickbait articles is at least second hand, usually some common knowledge or even urban legends. I'd rather have a google result page with only one result but that's the primary source, than a page full of different sites paraphrasing the original information and reshaping it such that it better fits the own site or provokes more clicks.

        • by allo ( 1728082 )

          Addendum: you may say "But what if there is no primary source?!"
          This sounds like a problem in the first moment, but if you look a primary sources, they are rarely put on sites that try to foremost get ad clicks, but rather on boring sites that convey the information straightforward without hyping it.

          News: Computers now faster
          Clickbait: Ten reasons why computers are now faster!
          Source: Intel's describes its new microarchitecture
          Summary: The newest architecture of Intel processors allows faster matrix multipli

    • Yet Google encourages the exact opoosite with Youtube. 8 minutes of drivel for your three word answer.
      • by allo ( 1728082 )

        Google is in a strange position there. They profit from the 8 minute video, but then put a generated summary below it, that can cut it short. And I think you can just put a Youtube Link in Gemini to get a full transcript.

  • by PPH ( 736903 ) on Friday July 31, 2026 @03:31PM (#66266706)

    ... your site. But I'm tired of being yelled at for turning on enhanced tracking protection. Not blocking adds. I see Retool in my sleep by now. Just asking "Please don't track me."

  • Now they seem to blindly trust AI.

    You should still be skeptical of Wikipedia, but at least you can see a citation...unless there is none or the citation is unavailable.

    • by allo ( 1728082 )

      Wikipedia is at least right most the time. The problem is they were trusting the first Google result, even when it was Infowars.

  • oh no! AI is coming for my click-bait articles!

    It's just another pebble on the "AI is good" side of the scale if you ask me.

  • by awwshit ( 6214476 ) on Friday July 31, 2026 @04:15PM (#66266764)

    I can't believe how many of those AI summaries link to Reddit. Then you follow the link and its some random people asking questions with no good responses.

    Seriously, follow those AI link they are garbage.

    The AI summary might limits clicks but that does not make it good or right.

    • by allo ( 1728082 )

      Unfortunately Reddit seems to have a trust bonus. Google clearly has some whitelist with sites like Wikipedia, Reddit, Stackexchange, etc. that they favor for queries that might have an answer there, drowning the sites that have a more in-depth answer on that topic.

  • The results that come first aren't there because they're useful, but because the site's owners have gamed the algorithm to come first, and then surrounded that algo-gaming content with as many tangibly related advertisements as possible in the hope of generating some cash.

    Having an LLM help parse through the bollocks has been a refreshing experience, honestly.

  • Scrape enough of other peoples work and content, then, don't LINK to it. No need to visit the sites made by creators. Not when Google can monetize it to everyone with advertising. If searchers go to the original sites, then Google might have to share a bit of ad revenue. (Can't let that happen again!) They do it with photos/images, news... virtually everything. And sadly, most folks are happy for the convenience. Saves them time.
  • I stopped using Google over a year ago, since first it delivers too many 'sponsored' links and the rest are links that it THINKS I want, instead of reading what I actually wrote.
    Then I would have to read all those results for the data i need.

    AI does all the searching, reading and extracting the data without any work from my side, it formats it exactly to my needs to paste into my research/projects/whatever.
    Google is dead to me.

  • by cfalcon ( 779563 ) on Saturday August 01, 2026 @09:04AM (#66267574)

    For several years now, any reasonable simple question has had Search-Engine-Optimized useless shit at the top of google- these days it generally can be counted on to fill the entire first page. Google helpfully highlights the relevant text by default so they spend money making sure google can't displau that text but instead displays an ad.

    I just googled:
    odds of rolling three sixes on three dice

    The AI has the 1/216 answer and the reasoning and the math and the links to the people on forums explaining all that. That is actually the search engine right there; it actually links me to people who have explained this to each other over the years.

    The top link is to reddit and claims the odds are "180 to where", and eventually has a bunch of links (reddit and google signed an agreement to stop other search engines from reading reddit well). If I click in there I will find the answer. But what if it wasn't?

    Under that is stackexchange, and all the wording is chosen so I can't see the answer. Clicking through the answer is present at least. Below that is a dice probability calculator that can make the answer, and something called thoughtco which has the answer in the summary. This is the first non-AI spot on the page with the answer.

    Overall the AI summary has been godsent, but if more time has been spent fighting SEO it would not be so needed.

  • If Google didn't give us AI summaries, they'd just be accelerating the AI agent marketplace. You'd wind up asking an agent what the answer is and it would go out and find it for you. And that traffic doesn't read ads, or doesn't respond to them anyway.

The young lady had an unusual list, Linked in part to a structural weakness. She set no preconditions.

Working...