Anchor Text Distribution

So the simple question is...Does Google take into account the anchor text from every single backlink from every single page for the overall anchor text distribution to determine if you need an over optimized anchor text penalty?

A bit more detailed version:

When I have 20 backlinks to my URL coming from one website but from different pages of that website does Google count the anchor text from each backlink when they are determining if you need a over optimized anchor text penalty.

Or for example would a site-wide backlink which turns out to be 1000 backlinks containing one anchor text cause an over optimized anchor text penalty?

I am trying to work out the best way to count my anchor text distribution. I found many tools and many different articles getting you to do different things, which all cause completely different results.

Some say only count one backlink from one domain. So if I had 10 backlinks to mywebsite.com from hiswebsite.com (but on multiple pages of hiswebsite.com) I should only count it once. This is how ahrefs works I think.
Another says I should count the anchor text from all 10 of those backlinks if they are on different pages.

I know that if Google finds multiple links on the same page pointing to the same place then they should only count the first backlink and it's anchor text. But then what if there are more on other pages?

What do you think?

Interesting. I also want to hear what pros will say about this. I have just started doing seo for my website so this kind of info is crucial for me to lessen the mistakes I could be doing in my seo process.

After doing some research on ahrefs I can see that some really big brands have a really bad percentage of over optimized anchors if you look at total referring pages, but then if you look at the anchors for the total referring domains instead their percentages look ok.

Considering these big brands have not been hit with any kind of penalty, I think we should be looking at our overall anchor text distribution by total referring domains rather than total referring pages.

It could be that site-wide links themselves give penalties due to them being non natural and not the fact that it skews the anchor text distribution.

If you have 10 backlinks from 10 different pages on the same website then Google has said that the link juice is split between all of them essentially making only 1 worth anything. So if that is the case I would imagine the same happens for the anchor text.

I think my new plan is to watch the anchor text distribution by looking at the anchors per referring domain, and then also keep an eye on big site-wide links which are no good anyway. Once that's done the anchor text distribution should be similar on both referring domains and referring pages with a natural link profile. Not many sites out there would link to another website multiple times with the same anchor text anyway.

I forgot to mention, If I look at my own site on ahrefs (which has currently got a definite algorithmic penguin penalty which I believe to be anchor text over optimization) then my anchor text percentages are absolutely fine when looking at total referring pages but when I look at referring domains it is obviously unnatural. Another reason I came to the conclusion above.

I would play it safe and modify your anchor text every time no matter what. So if your keyword was blue sneakers, make your anchor text best blue sneakers, top blue sneakers, blue sneakers on your feet, etc... That way you are 100% covered. Otherwise build a ton of brand links (url, click here, anything unlike your keyword but related) Then you can go out an be aggressive with your exact keyword anchor text. Hope that helps.

When using site wide links, it is recommended to use branded anchors or bare urls. Also, total referring domains/pages for your main anchor should ideally be below 10-20% for lasting results.