Rethinking Internal Linking Best Practices
Most SEOs understand that links are an official Google ranking factor.
It’s also generally understood that external links (links from other websites) are typically more powerful than internal links (links from your website.)
But let’s not underestimate the power of internal links. Because indeed, when used correctly, they can be very effective, as the data suggest.
For this study, we looked at 23 million internal links across 1,800 websites, representing approximately 520,000 individual URLs. We then compared these to data from Google Search Console to determine search clicks for each URL.
A few caveats about the numbers:
- Twenty-three million sounds like a lot of links, but it’s only a small portion of the trillions of links across the internet! A much larger study would provide greater accuracy.
- This is mostly a correlation study. Remember that correlation isn’t causation (though it can be a very strong hint!)
- We recommend you consider any conclusions drawn from these numbers directionally useful but not necessarily scientific truth.
Internal Links Associated With More Google Traffic
As you might suspect, pages with more incoming internal links tend to get more traffic.
For example, in our dataset, URLs with 0-4 internal links only saw two clicks on average from Google Search, while URLs with 40-44 internal links saw four times that many.
But then a strange thing happens.
After a URL receives about 45-50 internal links, the effect reverses. Google traffic begins to decline as the number of internal links rises.
What’s going on?
Diving into the numbers, the most obvious explanations seems to be explained by the existence of navigational/sitewide links vs. unique links in the body text. When a page has 50, 100, or 400+ internal links pointing at it, there’s a good chance these URLs are linked in the navigation from every page on the site.
With sitewide links, the dataset often shows URLs with less traffic.
Does this mean navigation links are less effective than other types of links? Not necessarily.
As the number of links per URL increased, our dataset grew very spikey. Observationally, large, high-traffic sites tended to do very well with navigational links. At the same time, small-to-medium sites seemed to have less success.
But then why are more links associated with more traffic, at least up to a certain point?
Perhaps looking at anchor text provides more insight…
Anchor Text Variations Correlated With Search Traffic
When talking about anchor text, we’re talking about the words that make up the clickable part of the link. The following are possible different types of anchor text (all pointing to this page.)
Pages can have only a few anchor text variations pointing at them, or sometimes many. Popular pages on the internet typically have a wide variety of anchor text variations from external links.
But is there any association between anchor text variety and traffic with internal links?
The relationship between anchor text variety of internal links and Google search clicks was so strong that we ran the data three times. Even after eliminating nearly all the outliers (close to 50% of all URLs), the numbers kept increasing.
The numbers kept increasing, but the data became less reliable after a certain point. While it’s common to see pages with one, five, or even ten anchor text variations pointed at them, only a tiny percentage of URLs have 25 or more anchor text variations from internal links.
Regardless, URLs with a larger number of anchor text variations from internal links are highly correlated with more Google search traffic.
Speculating, this may help explain why pages with more internal links see more search traffic, but only to a point. One theory may be the raw number of links may be less important than the uniqueness of the links.
Put another way, a sitewide link might appear on every page of your site (and in theory, it may pass a good amount of PageRank), but it can only ever have one anchor text associated with it. Even if that link appears on 500 different pages, in some ways, it might be considered a single editorial link.
Again, sitewide/navigation links seem to do better on large, popular sites, but the relationship between anchor text variation and traffic seems to hold across all sites, regardless of size or popularity.
Interestingly, Authority Hacker published an analysis of anchor text variety and Google search rankings. That study was very different from this one regarding data and methodology but with strikingly similar results. It seems anchor text variety is important for external and internal links both.
Naked URL Anchors and Google Traffic
Next, we wanted to look at specific types of anchor text links in our dataset.
One such type of link is when the anchor is simply a URL, sometimes known as a “Naked URL” anchor, like this: https://zyppy.com
Google’s SEO advises against using URLs as anchors, telling webmasters to avoid “Using the page’s URL as the anchor text in most cases.”
But does using URL anchors hurt your search traffic?
In our dataset, the answer was no harm at all.
Pages with URL anchors from internal links saw almost 50% more traffic than pages without URL anchors.
Speculating again, perhaps this has more to do with anchor text variety than any usability or other Google ranking consideration. Many SEOs are known to use naked URLs as part of their overall linking strategy.
To be fair, while you may see URL anchors everywhere, they actually aren’t that common. In our dataset, they represented less than 1% of all links.
Empty Anchor Texts and Google Search
Google also tells us to use descriptive words in our anchor text.
Despite this, links often contain no anchor text whatsoever. Image links—which use the alt attribute as anchor text—are particularly bad at this.
In our dataset, over 6% of all links contained no anchor text, but was it associated with fewer clicks? Here’s the answer.
Statistically, the existence of empty anchor text from internal made no difference whatsoever.
This could be because—as previously mentioned—empty anchors are most often associated with image links, and image links are often accompanied by regular text links, which presumably have anchors!
Exact Match Anchors and Google Search Traffic
All good SEOs advise using keywords in your anchor text, but there’s a lot of debate about using exact match anchors as part of your internal linking strategy.
Exact match anchors mean your anchor text “exactly” matches a top or desirable search phrase for the page in question, e.g., “best SEO blog” is an example of an accurate exact match anchor. 😉
For this analysis, we looked at pages with at least “one” exact match anchor from an internal link. The relationship was surprisingly strong.
Pages with at least one exact match anchor had at least five times more traffic than pages without.
This relationship held up no matter how many ways we sliced the data, how many outliers were limited, or how deep we looked into the dataset.
But remember, correlation isn’t causation. This might be explained if pages with exact match anchors are more clearly optimized to target specific terms.
Regardless, the evidence suggests that using at least some exact match anchors in your internal linking strategy may not be a bad idea.
How to Optimize Your Internal Links
While we can’t pretend to present a complete internal linking strategy based on this study, the data does hint at some possible best practices and optimizations.
- More internal links are associated with higher traffic, but only to a point.
- Sitewide/navigation links seem to have a powerful effect mostly on larger, high-authority sites. The effectiveness is less clear on small, lower-authority sites.
- Anchor text variety is highly correlated with higher search traffic.
- Naked URL anchors don’t seem to hurt and in fact, are associated with more traffic.
- At least some exact match anchors are associated with significantly higher traffic.