A page can easily have 10 000 times more incoming internal links than another page with the same type of content – two product pages for instance. It’s actually quite common. Such enormous disparity can’t be right, in terms of internal linking optimization.
Of course the value of each individual links varies, depending on how much “link juice” the source page has to transfer, and depending on the semantics it transmits through the anchor text. But the number of incoming links remains a pretty good indicator to compare different areas in your website.
Very few links means very few visits or none, and a high number of links means much higher chances of getting organic traffic.
That’s why the Botify Log Analyzer crawl report shows a graph with both the average number of incoming links per page and the number of organic visits per active page. You can find this graph in the ‘conversion’ page:
This example shows that visits per active page loosely go up and down with the number of incoming links.
Here is another example: the graph below is from a website with editorial content, it shows tag pages – they present lists of articles on a given topic. They are categorized as ‘first page’ vs ‘pagination” (page 2 etc.).
The first pages receive virtually all links, and all organic traffic.
Note that in this case, this is exactly what we want: the first page of a list is the only target for organic traffic. If a user searches for “crisis in the middle east” on Google, the first page of the list of articles with this tag is the appropriate answer to the user’s query – not page 2 or page 10.
So we need to make sure that most of your internal linking goes to pages that are targets for organic traffic, and that all target pages receive a fair amount.
You website really needs internal linking optimization if:
1) A significant portion of your internal linking goes to pages that are not targets for organic traffic
2) Some pages that are targets for organic traffic receive a single link, or just a few
3) Within a category, some pages receive A LOT more links than most
This is easy to see in a graph that shows the average number of incoming links per category (internal linking section in the Botify Logs Analyzer crawl report):
Here, we are looking at averages values, which provide a bird’s eye view and don’t show disparity within a given category of page.
When we want to identify more specific trends, decile graphs are extremely useful.
About incoming links decile graphs: each point indicates the average number of incoming links for a 10% chunk of the overall volume of pages. For instance, let’s say there are 5 000 pages in the category from the graph below. The last point means that the 500 pages with the highest number of incoming links have on average 127K links per page.
In this example, the point we want to make is that 20% of tag pages receive a single link, another 10% receives only 2. Clearly, these are at a disadvantage when it comes to ranking.
In this particular example, this even questions whether the pages in the first two deciles should exist. As these are pages with lists of articles, each tag page is supposed to receive a link from every article with that tag. This means these are virtually empty tags. This graph does not show how many there are. To find out, we would need to look at the data tables that are delivered separately.
The third common scenario is the following: in a category of pages that are identified as targets for organic traffic, most incoming links are concentrated on a very small percentage of pages.
Of course some variation in the number of incoming links is expected: for example popular products, special offers, and other pages we want to promote receive more links. Navigation pages also naturally receive more or less links depending on their position in the site structure: a ‘/books/travel/asia/thailand/’ page will receive less links than a higher level page such as /books/fiction/, because of mechanisms such as breadcrumbs, which creates a link to the page from each product in the section.
But when disparity comes to the point where there are 1 000 times, 10 000 times, or perhaps 100 000 more links to one page than to the next, redistribution is clearly needed. Receiving a bit less won’t make the difference for these heavily linked pages. Receiving a bit more will make the difference for the other ones.
In this graph for instance, 90% of type B articles receive 1 000 to 30 000 times less links than the top 10%. Differences are also huge for the other types of articles.
There is a problem when there are many orders of magnitude between the lower and the higher end, and a shape that is almost flat, and raises almost vertically for the last decile.
This typically happens when some links are repeated on every page in a section, or worse, site-wide : a ‘top products’ block for instance. A much better approach is to implement some ‘top products’ per section : if the website sells books, we can have top novels in the novels section, top comics in the comics section etc. down to subsections. These links will also be more valuable for SEO because they are more valuable for the user, too (contextual links).
Another effective way to distribute links is to implement a ‘You may also be interested in ‚Ä¶.’ box for product-to-product or article to-article links with different suggestions on every page.