SEO networks: types of auxiliary sites and link architectures

Almost any page found on the web can become a zaplecze (an SEO support network). As a conscious or unconscious action, it is an important support for the activities of a company or an individual in these areas:

• Promotional, by highlighting its presence on the web
• Image, building brand recognition, trust and consumer appeal
• Educational, by passing on knowledge relevant to the audience
• Informational, announcing news, supporting knowledge and education
• Sales, monetising the work put into building the network and raising the return on investment (ROI)

Building such support, from an empirical point of view and considering the activities above, is part of the long-term process of creating a website and can decide whether a venture succeeds. The breadth and reach of getting to users depends on the work of marketers: intensifying certain activities, avoiding mistakes, or carrying out actions that are not allowed. Relying on so-called “proven actions” when building link networks usually does not bring the expected effects, because the environment in which the network can be built keeps changing, and so do the forms of presentation. It is an unavoidable technological paradox in such a civilised society, in which one has to follow trends constantly and adapt to them flexibly, building networks in many places and areas of life. Depending on where hyperlinks appear and where they come from, we distinguish two types of networks: internal and external.

#Internal network within the site

One of the first stages of building a link network is preparing the target page, the one we build the network for, for crawlers. This is done by applying SEO guidelines on how the source code is built.

#Internal links in the content

Site code optimised for correct movement between pages should have a systematic set of internal links. On each subpage, in the page content, there should be hyperlinks pointing to keywords relevant to building the external network. The most frequent mistake is internal linking to pages of little relevance from a search marketing point of view. From the start of building the network, it is worth defining the most important strategic subpages that need promotion and optimising them for indexing. Links within the site usually gain weight in the analysis made by the search engine indexer, because they sit on a page thematically related to the hyperlink, which in a network building strategy is an important factor for reaching higher positions in the SERP rankings[1]. Links pointing to the specific subpages that matter most to us, when their hyperlinks are systematic, assure crawlers that each time the hyperlink leads to the same page. The form of the direct hyperlink to the page should be decided in advance and applied consistently in internal and external link building. A danger at this point is building the link in different ways, for example:

http://www.site.com/subpage.php
http://site.com/subpage.php
http://site.com/subpage.php&id=999
http://www.site.com/subpage

Although the intentions of the person using such forms were the same, referring to a given page in these different ways tells indexers to go to 4 different pages. Because of that, the target page does not actually gain support from the internal network.

#Text menu and image menu

A text menu supported by title attributes, or an image menu containing alt attributes, makes it easier for crawlers to reach the related subpages of the site.

#HTML sitemap and XML sitemap

In a strategy for building an internal network, the sitemap plays an important role. Done correctly, it should show all subpages of the site that crawlers should find, in a hierarchical and systematic way. This kind of page listing helps users or crawlers lost in the structure to find out, after a short contact with the page (before they leave it), where the target page is[2]. Because of their function, two types are distinguished: sitemaps in HTML and sitemaps in XML format.

The first present the hierarchical structure of the site as a list on one of the subpages, covering only the subpages the site owner deliberately lets search engine crawlers enter.

XML sitemaps are created as a systematic source of information for crawlers, and their use for people is minimal. Like the former, these maps contain the hierarchy of the site structure, extended with information about the creation date and modification date of a given subpage, and a suggestion for crawlers about how often specific subpages should be visited, so that the site does not use up unnecessary traffic.

Here you should make sure that the crawler, and later the spider, collects for the indexer all target pages that should be indexed. However, frequent 404 errors, an inconsistent form of internal links between pages, cloaking, orphan pages, or a system engine serving content based on cookie sessions can make this harder.

Traps a spider can meet when moving between the pages of a site

#Breadcrumb navigation

Breadcrumb navigation is considered the second alternative to the main menu, or the third after the sitemap, for navigating a site. It is usually located above the page text and helps users find their way around the site. Its advantage is quick access to higher-level pages when a user or crawler gets lost[4]. Together with the sitemap, breadcrumbs are an important cognitive element for crawlers, because they show the hierarchy of the site and how its subpages depend on each other.

#How Google spam policies evaluate types of SEO networks

The split into internal and external networks dates from 2008. The Google Search spam policies describe practices today, not categories of pages. Each type in this list therefore has to be assessed separately: the same technique, for example a sitemap or a blog article, can be normal site structure or a violation, depending on why it was created. Google detects violations with automated systems and, when needed, with human review, and sites that violate the policies may rank lower or not appear in results at all. None of the types described below is exempt from this assessment.

The definition of a support network, how a link is built, the rel attributes and the examples of link spam are covered in the article what an SEO link network is. A plan from 2009, assessed point by point against the current policies, is in the link building case study for the furniture industry. Here is the perspective of the types: what from this article’s list Google treats as site structure and what it treats as spam.

#When internal linking practices violate Google guidelines

Internal links, a menu, a sitemap and breadcrumbs are elements Google expects on any site. Google’s documentation on crawlable links recommends that every important subpage has a link from at least one other page on the site and that links to related content are in context. The same page adds that there is no ideal number of links per page, but if the number seems too high, it probably is.

The internal network enters spam territory through three policies:

• Hidden text and links. The examples include white text on a white background, text hidden behind an image, text positioned off screen with CSS, font size or opacity set to 0, and a link attached to a single small character, for example a hyphen in the middle of a paragraph. Internal links visible to the crawler and invisible to the reader fit this definition.
• Keyword stuffing. The policy refers to filling a page with keywords or numbers to manipulate rankings. The links documentation also advises against stuffing anchor text with keywords and placing links one after another in a chain.
• Cloaking. The 2008 text names cloaking as a spider trap. In the current policies it is a violation of its own: showing different content to users and to search engines, for example inserting text or keywords only when a crawler requests the page.

#How Google picks one address among several forms of the same subpage

The four forms in the example above are, for a search engine, four URLs. Google describes canonicalization of duplicate URLs as choosing one of them: the one that should appear in results and gather the signals of similar or duplicate pages. The documentation lists the methods starting from the strongest signal:

The methods can be combined and work better together. Between the HTTP and the HTTPS version, Google prefers HTTPS, unless there are issues such as an invalid SSL certificate. The 2008 recommendation, to fix one form of the link in advance and use it everywhere, still holds. Canonicalization tidies up whatever still appears in another form, for example an address with a campaign tracking parameter.

#How Google reads links in text and image menus

Google’s crawler only reliably extracts an address from an <a> element with an href attribute. Constructions like <span href="..."> or <a onclick="..."> are marked in the documentation as not recommended: Google may try to interpret them, but does not guarantee it. A menu built only with scripts therefore stops working as an internal network, even if it works for the user.

In an image menu, the anchor text is the alt attribute of the image used as the link. An empty alt on such a link appears in the documentation as a bad example. Good anchor text is descriptive, concise and related to the page it is on. Google suggests a simple test: read only the anchor text, out of context, and check whether it still says where it leads.

#What Google does with the modification date and frequency in an XML sitemap

The 2008 description of the XML sitemap assumed that a hint about visit frequency saved traffic. According to the guide to building sitemaps, Google today ignores the <priority> and <changefreq> values. It uses <lastmod> when the value is consistently and verifiably accurate. The date should reflect the last significant change to the page: in the main content, the structured data or the links, and not, for example, updating the year in the copyright notice. A single sitemap can be at most 50 MB uncompressed or 50,000 URLs. Larger sites split their URLs into several sitemaps and gather them in a sitemap index file.

The XML sitemap remains a weak canonical signal (table above) and does not replace links in the content and the menu. A subpage that is only in the sitemap, without any internal link, is still an orphan page.

#Why content that depends on a cookie session does not get indexed

Content served only in a cookie-based session was on the 2008 list of spider traps. Today, the documentation on fixing JavaScript problems in Google Search explains the reason precisely. Google’s Web Rendering Service (WRS) does not keep state across page loads: Local Storage, Session Storage and HTTP cookies are cleared on each load. Every subpage is, for the crawler, a first visit. If the target page only shows the right content after saving a choice in a cookie, the crawler sees the version without that choice.

#How to mark up breadcrumb navigation for Google

Breadcrumb navigation shows the position of a page in the site hierarchy. Google uses breadcrumb markup in the body of a page to categorise the information from that page in search results (breadcrumb structured data). A BreadcrumbList must contain at least two ListItem elements in the itemListElement property, and each element points to a page in the item property. The HTML breadcrumbs and their structured data should describe the same path the user sees.

#External network outside the site

Building an external network is more than just inserting hyperlinks to the target page into all kinds of materials available on the web and outside it. It is also a philosophy of running a company, a melting pot of PR, marketing, visual identity, brand identification and search marketing. This is the area that develops most dynamically, where a lack of full commitment and consistency, combined with carelessness about mistakes, can bring not only lower interest in the target page but its complete exclusion.

• Information pages
• Blogs and presell pages
• Directories (among others the most popular Polish ones of the time, QlWeb, mini)
• Discussion forums
• Copies of the DMOZ directory, Gadu-Gadu status descriptions, song lyrics, wikis, etc.
• Offline/BTL/ATL
• Intranet
• Other areas of life

#Which external network types Google describes as spam

The row with copies is the one that changed most. In 2008, a copy of the DMOZ directory or a collection of song lyrics gave cheap content for links. The current policies call that scraping, and gathering in bulk pages made of feeds, search results or fragments of other sites is also scaled content abuse. The assessment of information pages and blogs depends on purpose, because the policies ask why the pages exist: a blog network created for links falls under link spam and scaled content abuse, and the blog format itself decides nothing here.

#Why check the internal network before external link building

The internal part of the network depends only on the site owner, so that is where it makes sense to start. Before looking for links on blogs, directories or forums, confirm that the target page gets all the support the site itself can give it.

First, choose a single address form for each subpage and use it in all internal links. Writing the same address in different ways leads indexers to different pages and scatters that support.

Next, review the menu. A text menu with title attributes, or an image menu with alt attributes, lets crawlers reach related subpages.

Then keep two sitemaps: one in HTML for users, with only the pages open to crawlers, and one in XML with creation and modification dates.

Finally, look for what interrupts the spider’s path: 404 errors, orphan pages, cloaking and content served based on cookie sessions. Breadcrumbs complete the structure and give the user a quick route to higher-level pages.

With this foundation in order, external links start reinforcing the right page.

[1] Jeffrey Smith, October 12, 2007, seodesignsolutions.com [2] Jakob Nielsen, Site Map Usability [3] Elliance SEO Infographics [4] Jakob Nielsen, Breadcrumb Navigation Increasingly Useful

On the delivery side we cover this topic under GEO and LLMO.