Google says that “links are less essential” – we looked at 1,000,000 SERPs, and our data agree and disagree

Google said that the links are less essential than before, and most SEO will accept it as true. Our findings indicate that this is generally true, but for some types of queries it combines more.

I still think that links should be part of your SEO strategy, but you should not be all your strategy. Some time ago I answered the question: “Do the links still matter to the rankings?” Rejecting all links to several blogs. Movement and rankings have fallen, so yes, the links still matter.

This time I was curious whether the meaning of the links changed over time. I was also curious if we could divide the data to see when the links are more essential and when they are less essential.

We will go (after great thanks to our data scientist XIBEIJIA Guan for performing all the demanding parts of this study and for our CMO Tim Soulo for its contribution).

We chose 1,000,000 keywords with the highest amount of search from the USA and calculated Spearman correlation Between the SERPS ranking (20 best) and various SEO indicators, including:

• Domain assessment: 0.131
• Number of feedback: 0.248
• Number of the following feedback: 0.242
• Number of refdomen: 0.255
• Number of the following repairs: 0.250
• Number of internal links: 0.117
• Number of internal industries: 0.093
• Number of external critical: 0.083
• If the keywords present in the URL: 0.034
• URL length: -0.107

I want to give all normal warnings such as Correlation ≠ Cause. These are generally considered feeble correlations, and even very indigent correlations on a special -manana scale, but this does not mean that these things are not essential. There are many ranking factors, and links play a role in your rankings.

Here are some charts that visually show data on essential indicators. If you see box plots for the first time, here’s how to read them:

Diminutive lines on the edges represent minimal and maximum values. And 50% of all values ​​fall in the distinguished areas. The line in this area is a median value.

Basically better link indicators = higher rankings.

If we break them according to the search volume, we see that the links better correlate the search with higher search volumes.

I see 2 possible explanations and it is probably in both truth:

1. These queries are probably more competitiveAnd you may need links to move the needle for these competitive queries.
2. These pages obtain a larger exposure due to the higher volumeSo, there are more links with higher rankings.

External cables correlated better than incoming internal links. I was surprised how gigantic the difference was here. Links from other sites are probably more essential or seen as more reliable than links on your own site.

The links correlate better with brand searches than a non-marry search. But I don’t want people to interpret it as links are more essential for the brands. In my opinion, the right way to interpret this is that due to the popularity of many brands, they usually have more links. Their links and rankings No. 1 distorted these results.

The other side of the equation is that the connections correlate slightly less with the brands of the query of the above chart, which would somewhat lower these numbers. This context is essential for the historical comparison below.

We did a study in 2019 that looked at some of the same correlations.

The methodology was slightly different. It was low volume (2000-5,000), not branded queries. I cannot make a thorough comparison, but the correlations were higher at the time:

• Cufflinks: 0.27
• The following links: 0.25
• Recommending domains: 0.29
• After dismissal of domains: 0.26

Current correlations for these indicators and similar queries are 0.22–0.24 and may be slightly lower than for non -brands, as I mentioned earlier.

This is consistent with the fact that Google comments on links become less essential and how SEO perceives them. However, one of the reasons I wanted to conduct this study is to see how it changes in the next few years. As the AI ​​content increases, I suspect that Google may finally rely on links or other signals and may start higher.

Links and RDS correlated at 0.33 for local queries, which are much higher than what we saw, and even higher than what we saw at all years ago. I suspect that because the content on the pages of local service companies is very similar, the links are a good distinguishing feature to see which company is popular.

I want to point out that internal links seem to be more essential for local queries. I suspect that for these queries there are less external link signals, so additional signals from internal links become more essential.

If we look at the data according to the search intention, navigation queries have the highest correlation. They are mainly branded and, as I said, brands have more links.

It is more compelling that the feedback for information inquiries seem to have a higher correlation for rankings than commercial and transactional queries.

I suspect that because competition levels for these conditions are high, Google can be considered a good signal. It may also be that better content obtains more links over time.

Final thoughts

Links are part of a larger ranking puzzle, but you still have to do other things, such as creating great content. Where Google has more signals, they can rely on one signal. Where they have fewer signals, they can rely more on such things as links. In areas with high competition, in which the content can be robust, they can also rely on link signals as more distinguishing features.

I think that companies, especially local companies, should probably focus on at least fundamental links. Larger companies probably already have marketing efforts and content that brings them links, so links may not be so priority. However, when pages for immense companies require pushing, I saw that even internal links give a nice impulse.