Google recently hosted an event that invited a small group of website owners who were highly impressed with its helpful content system.
I’ll summarize some of the great blog posts and tweet threads written on this topic. Then I want to think about what we can learn from what Google representatives say.
The helpful content system started as a machine learning system that learned from examples of helpful and unhelpful results. Google’s systems have evolved to the point where they can now use a variety of signals to predict what content a searcher is likely to find helpful.
In the end I would like to think about what to do if you are affected. While I think some Although sites can recover, I think most severely affected sites are unlikely to see significant improvement. I know this is discouraging. I know so many good, hardworking people who are stuck and don’t know what to do after being disrupted by this update. You are not alone. In fact, as AI advances, the entire world is changing and changing the way knowledge is shared.
But first let me say a word about Google.
I’m not happy with how Google handled HCU
I understand the need for a helpful content system. Overall I think it was good for the web. The vast majority of affected sites were sites created solely for the purpose of making money, rather than to truly inform and help an audience. However, recent Google updates have destroyed the existence of some really good people and websites. And as we’ll see in this article, Google doesn’t know how to fix this.
I’ll end this article with thoughts on how to decide whether to proceed if you’re concerned about the helpful content system. I wish Google would release something to help site owners understand where search is going and why HCU needed to happen.
As I write this, ChatGPT has just released its search engine. The way people search for information online is changing, which means traditional models of making money online are changing too. We are in the midst of a significant shift in the way knowledge is shared.
Observations from participants
Several of the participants said they were told their content was good, but Google’s systems couldn’t detect it. It was disheartening to hear that there was little hope of a speedy recovery.
I drank the Kool-Aid at the Google Web Creator Summit 2024 -Mike Hardaker
- Mike made $250,000 last year. This year he is eating at the food bank and is barely making ends meet.
- Danny Sullivan told him, “My content wasn’t the problem.”
- Pandu Nayak: “I suspect that you guys create a lot of great content that we don’t make available to our users, but unfortunately I can’t give you any guarantees. We’re focused on things for our users, that won’t change.”
- Danny Sullivan: “The entire format of search results has changed.”
- Regarding Forbes: “Google’s signals may be better at figuring out current authority topics that the company is currently struggling with.” I think that’s interesting considering the blog post I just wrote see how Forbes performs lost keyword rankings to some real topic experts.
I attended Google’s Creator Conversation event and it turned into a funeral -Joshua Tyler
- Danny said there was nothing wrong with their sites.
- There is no site-wide classifier. Individual pages are rated at page level.
Kim Snaith
https://x.com/ichangedmyname/status/1851652836917985627
- They received no answers.
- Danny: “There’s nothing wrong with your websites, we are.”
- Someone asked Danny what he would do if a site was affected by HCU, and he replied, “In the meantime, I would do something else.”
Morgan
https://x.com/CharlestonCraft/status/1851465285557207228
- Google employees denied there was a problem and don’t know how to fix it.
- She said: “If you’ve been hit by HCU, don’t expect recovery any time soon.” “Move on.”
- An update is coming “very soon”. However, don’t expect any recovery until perhaps a few updates later.
https://x.com/CharlestonCraft/status/1851643609549787382
What can we learn from this?
Google invited these creators because they had good websites. That’s why they were invited – so Google can learn from them and potentially improve its detection of other types of helpful content.
Why were they so influenced by HCU? And why does Google say not to expect a recovery anytime soon?
Here is the one I found this part the most interesting:
Google has asked several times for ideas on how to distinguish between a spam site and a good site. That’s interesting.
The helpful content system was a machine learning system
We know that the helpful content system was a machine learning (AI) system. Machine learning systems are trained by seeing good and bad examples. They then determine which features they can consider and how much weight to give them to predict whether an unseen example is good or bad. In this case, the examples are helpful and unhelpful results.
We know one source of these examples of helpful and unhelpful results: the quality raters. That’s in the Guidelines for Evaluators:
The Statement by Douglas Oard from the DOJ vs. Google test report tells us about another source – live experiments.
Google can learn a lot about what types of pages are helpful to searchers by studying what people click and interact with. These sites are more likely to be helpful than not helpful.
If your website has been compromised by the original helpful content system, it means that this is the case is more consistent with examples of unhelpful content than with what is considered helpful.
There is now a new system
Google’s systems have changed since the original helpful content system was introduced. The Original documentation Released with the March 2024 core update, Google tells us that the core systems now use a Variety of signals and systems to understand the usefulness of individual pages.
I strongly suspect that the user data and reviewer input are of high quality was Originally, the system was used extensively from 2022 to early 2024 with the original helpful content system. During these two years, the systems learned by looking at examples provided either through reviews from quality raters or through live experiments monitoring user clicks to predict whether content is likely to be helpful or not.
We also have an example in Oard’s statement that shows how user data may become less necessary as AI capabilities improve.
They talked about how, in the early days, Google used user input to learn spelling corrections.
The summary judgment The course of the trial tells us how AI ultimately made it less necessary to use user data.
I believe that as of the March 2024 core update, Google made a shift so that its AI systems could use far more on-site signals than user signals to predict what content was likely to be helpful and what was not. After two years of experimenting with the helpful content system, the AI systems had learned a lot about the characteristics of helpful and unhelpful websites and were able to do most of the work themselves.
As Google announced, the March core update marked one Evolution in the way they identify the usefulness of content.
The unfortunate thing is that the system is not perfect. Some decent websites were unfairly labeled as unhelpful by this system, with devastating consequences.
What signals are associated with (un)helpfulness?
We have no way of knowing exactly which signals are being measured and how much weight is being given to them. I doubt that anyone who works for Google even knows.
Here are some thoughts based on years of studying websites that Google’s systems deem low-quality. These are some traits that I think are consistent with a lack of helpfulness. If you are actually still working on recovery, these may help you.
Is it difficult for the searcher to find what they read on your site?
Don’t let your readers scroll past endless chatter and unnecessary words. Place the part that matches user intent high in your content. Or at least give the reader something in the early text that makes them feel like they’ve landed in a place where their needs are being met. Use headings to make it easier for them to skim and navigate.
Many websites affected by HCU forced the reader to scroll through several pages to get to the part that interests them.
Are you relying too much on people-also-ask research?
While it makes sense to include related questions in your content, if the majority of your content answers the PAA questions, you’re probably rewriting content that Google already knows about.
This type of content performed well before HCU as it looked good to vector search relevance systems. However, this type of content often contains a lot of information that real people tend to skip over.
Which leads me to the next point.
Do you have original and insightful information?
Here’s an exercise to try. Search for a keyword you want your page to rank for. Think about what the searcher wants to find or achieve. Look at the first five results and write down three to five things a seeker would learn to achieve their intent. Now open your page and start reading from the top. How far do you have to read before you get to something interesting and important that hasn’t already been covered?
Are you using too many ads?
It’s perfectly fine to display ads on your website. However, we must remember that users do not like intrusive advertising. Large brands often get away with more advertising harassment than small providers. If your ads cause searchers to click away, this can send similar signals to Google as websites that exist solely to make money.
Do you lack relevant specialist knowledge/experience?
This particularly applies to informational content. Here is another test you can do. Search for a keyword you want to rank for and look at the experience of the websites ranked by Google. If you are behind the websites of a real store, a proven expert, or well-known authority sites, it will be difficult to beat them unless you can produce content that people actually want to read, even if they have already seen those results .
Do you have an audience outside of search?
Google emphasizes the need for human-centered content. The systems seem to favor brands that people know. The Guidelines for Evaluators Talk in detail about your reputation and being known and recommended by others. If you’re not sure if this is an issue for you, I recommend reading the guidelines and looking at all the places where “reputation” and “known” are mentioned.
or… it is possible that…
There is nothing specifically wrong with your content
You may have good content, but Google’s systems are not yet advanced enough to recognize it correctly. You may have decent content, but not content that really stands out from the sites you’re competing with. I know this is frustrating. My condolences go out to the many entrepreneurs who have been devastated by these changes. There is a change that will improve Google. But it sounds like that won’t be the case any time soon.
Is recovery possible?
Some of the creators who came to the summit expressed hope that they could eventually recover. But the general impression was that they probably not.
It sounds like Google is improving the system. But if they do, then so be it It’s unlikely to help in the short term.
It is possible to recover. Here’s an example of a site that recovered after a HCU hit in September.
This is not a customer, but they were one Newsletter subscriber for years. Looking at the old version of their website I can see that they have made drastic changes:
- Completely new, much more modern website design
- comprehensive and incredibly helpful buying guides
- Great use of headings to orient the user
- At the top of the page, user intent is fulfilled
- unique images that show real people actually using the product they are reviewing
- helpful videos in high quality
Here’s another one, although it’s questionable whether HCU’s website was affected or perhaps it was the previous core update from August. Their GSC data shows a significant decline as of September 14, 2023, so I believe the helpful content system influenced them. These people were customers years ago. I can’t take credit for the improvements they’ve made, but I know they read my thoughts in my newsletter and are working hard to create a helpful website.
They’ve added new pages that have done a really good job of answering their audience’s top questions. It should be noted that they serve a very specific audience. Not only do they write content to rank in search, but they actually help people in their industry. They have also added quizzes that really help the user answer the question they were looking for on the site.
I show these to say that Is possible to recover. But unfortunately it is The vast majority of the hundreds of sites I’ve seen affected by HCU have shown no signs of recovery Or show a pattern like this, where the August core update started to help and then the site started to decline again.
Should you move on and work towards recovery?
Here are some questions to consider that may be helpful.
1) How much of your success so far is due to your SEO skills? If SEO was your drive, you may find that old tactics no longer work.
2) How original and insightful is your content? Do you have competition? Can you be significantly better than them?
3) Do you have an audience outside of search? If not, you can work towards becoming a brand – by getting your name out there, doing podcasts and interviews, and more. If you don’t want to do that, recovery may be difficult.
4) As AI answers and language modeling tools like ChatGPT and Gemini get better at answering questions, Are you able to create a resource that people are actually looking for? also out with these available tools?
Here is a GPT I created that you can use Use ChatGPT to brainstorm whether recovery is possible. You can find a bunch of my advice and more in the knowledge base. Of course, I wouldn’t rely entirely on this tool… but brainstorming with it might help you clear things up.
The search is changing. The world is changing.
Search is changing really quickly. I believe that the sites affected by the HCU represent the beginning of a significant change in the level of civilization. We are already seeing news publishers close. Many companies that rely heavily on organic search are struggling.
My husband has a saying. At first I didn’t like it, but now I’m starting to understand it: “It’s happening for you, not for you.” For many of you reading this, I think it might be time to look for other options, Using your talents and skills. There will soon be an incredible demand for people who understand how LLM tools like ChatGPT and Gemini work. That’s why I created a new website to help people Learn more about AI. Or maybe there is another path waiting for you.
In 1900 41% of the workforce was employed in agriculture. The industrial revolution changed that. We are at the beginning of a new revolution as AI changes the way we live.
If you decide to further improve your website, there’s probably one step you’ll need to take some important Google updates before the systems recognize your hard work.
I would like to thank the YouTubers who took the time to speak to Google. Creators are incredible people and I have great hope for their future.
Mary