Traffic Drop Due to Google Search Better Matching

1. 94

   eugbugposted 20 months ago

   Do you think the drop may be partially due to how the SERPS match search terms more closely? For instance my "How to Sow Seeds" guide is still in first position for the search term "How to sow seeds", but if I search for sowing seeds, it's way down the list, although it used to be in first or second position before. Traffic is down around 70% compared to last year and 80% compared to 2021. I notice this for other articles too. If the search term doesn't match any of the words in the title, the article doesn't turn up, even if the term match H2 titles in the article. Anyway I'm going to change the title again, nack to what it was in 2021 to see whether it makes a difference, even though I have first place for the existing one.

   replyreport
   1. 86

      tsmogposted 20 months agoin reply to this

      No expert, here, but a curious sort. From the little, I just learned one must consider the impact of AI with ranking articles of a search query. Ignorant I discovered two articles sharing about AI with Google algorithms.
      
      The first article I read had short explanations of the different AI used today. They share about:
      
      RankBrain introduced 2015
      Neural Matching was introduced 2018
      BERT, Bidirectional Encoder Representations from Transformers introduced in 2019
      MUM, Multitask Unified Model launched 2021
      
      The article is How Google uses artificial intelligence In Google Search by Search Engine Land.
      https://searchengineland.com/how-google … rch-379746
      
      My curiosity led to an article explaining RankBrain opening my eyes to a new way for me how they connect dots to give querie's ranked articles. It is pretty thorough in explaining it and technical too. One short quote that caught my attention is; This is to say, it took it from “reading” literal characters, and instead “seeing” the entity they represented. I thought about that with what you pondered with "How to sow seeds" vs. sowing seeds. 
      
      The article is A Complete Guide to the Google RankBrain Algorithm by Search Engine Journal.
      https://www.searchenginejournal.com/goo … rankbrain/

      replyreport
2. 97

   PaulGoodman67posted 20 months ago

   I think, if anything, the algorithm has gone the other way as it's become more sophisticated. That's to say, in the early days, it was very much focused on exact matches. However, they've factored in more considerations and it's become more complicated over time.
   
   When a site gets hit, the first thing that can be lost by articles is the related keywords traffic. Essentially, an article can retain ranking for its primary keyword but lose ground for the related ones. It sounds like that's what you're describing to me.
   
   If I'm right, changing the title probably won't help you, unfortunately. It may even make things worse, as you could lose the #1 ranking for your primary keyword.

   replyreport
   1. 94

      eugbugposted 20 months agoin reply to this

      Yes, it probably is. My articles used to turn up for a variety of search terms. Now I have to type nearly the whole title for them to be listed.

      replyreport
      1. 97

         PaulGoodman67posted 20 months agoin reply to this

         HP's reputation problems look like they could get worse with AI coming along. Their problem is that they haven't really got the resources to fact-check articles with the necessary depth required.
         
         It's getting easier for anyone to put together great-looking articles that seem to contain feasible info but actually contains a lot of BS. If it's written on a specialist subject, then it really takes an expert in that field to see the problems.
         
         "Artificial Intelligence" is a misleading term, in some ways, because it doesn't really know or understand like a human, it's just gathering and manipulating info.
         
         I can see Google increasingly favoring specialist websites in many cases, rather than generalist ones like HP, because of the fact-checking capabilities. I hope I'm wrong.
         
         At the very least, I think HP will have to change in a major way how it does things.
         
         On a larger scale, it's not clear how AI will affect the whole search engine field itself. We may end up with a world where people ask AI a question and it just generates an answer. It's not clear how search engines would fit into that world.

         replyreport
         1. 94

            eugbugposted 20 months agoin reply to this

            Yes, I guess what Google may really want is a scenario where they have an absolute monopoly on providing answers. So Lamda or whatever their chat entity is called just answers reader's questions and SERPS are made redundant. Then they'll be able to make more money on ads on the AI answer pages.

            replyreport
            1. 97

               PaulGoodman67posted 20 months agoin reply to this

               I feel Google will probably do their best to defend their old article-based search engine territory while simultaneously trying to win the new AI war against competitors.
               
               I kind of wonder if Google's "reputation" updates are motivated at least in part because they envision the internet being flooded with increasing amounts of AI-generated material, which appears feasible but contains misleading info or falsities.
               
               I feel HP may have to rely increasingly on a pool of trusted writers, rather than be quite as open. Judging each article individually regardless of the author is problematic. That said, they do rely on "fresh blood" to generate content.
               
               In many ways, HP have been drifting away from the open platform idea for years. When I started, they published virtually anything by anyone. Nowadays, it's become more like traditional publishing, with editors acting as gatekeepers, suggesting changes, and altering the articles.
               
               It could be argued that AI isn't a paradigm shift as, in the past, some online writers were always able to dress up badly researched material as an informed article.
               
               However, the difference with AI is that it makes the process way faster and more efficient. Articles can be created almost instantaneously, and as time goes on, they will appear more and more convincing.

               replyreport
3. 97

   alexadryposted 19 months ago

   I have been finding the opposite. Can't recall what query I exactly searched for, but if I had an article on a specific detailed topic, search results showed, at the very top, all generalized articles, while my detailed article, which fit the search query sentence almost perfectly, was like at the 15th position using Google's scrolling feature.
   
   I don't know what's going on with Google, but I have also found an article of 420 words outranking my article of 1,200 and the author was just a random name with no profile!
   
   I am going to start updating my articles soon to see if it makes a difference. Traffic is going down again on my side after traffic appeared to be recuperating a bit.

   replyreport