Everything You Wanted to Know About Inbound External Links
Tom Schmitz Jul 31 2008
Today, a colleague asked me what I know about
external links. Rather than simply e-mail her back I decided to sit down and answer this like I might if I were taking a college course exam. In other words a time limited brain dump. Let’s see how much I remember. :)
A Little Foundation
- InLink = Inbound Link – A link from another page or document on the
Web to the page being examined.
- OutLink = Outbound Link – A link from the page being examined to another
page or document on the Web.
- Internal Links – Links between pages or documents within the
- External Links – Links between different domains.
- If a web page links to another web page more than once it is only counted as
Most often when we discuss links we mean external links.
Two Reasons Links Are Important
- Links indicate trust and authority
- The more pages that link to a web page, the more trusted that web page becomes,
especially when those links are from other web sites. (Quantity)
- The more search engines trust pages that link to a web page, the more trusted
that web page becomes, especially when those links are from other web sites. (Quality)
- The more domains that link to a domain, the more search engines will trust
the domain receiving those links. (Quantity)
- The more search engines trust domains that link to a domain, the more search
engines will trust the domain receiving those links. (Quality)
- The longer a link exists the more trusted it becomes. (Quality)
- The more pages that link to a web page, the more trusted that web page becomes,
A page receiving 1,000 low quality links may receive the same amount of trust
from the search engines as a page receiving 10 high quality links.
- Links provide context and relevance. - Words in links matter. If 1,000 pages link to a page with the anchor text
or link text, “iPod listening station reviews,” the search engines will thing that
the page being linked to is relevant for iPod listening station reviews. (The search
engines will want to confirm this by checking the contents of the page, but in the
past, pages routinely ranked for search engine queries where the only place the
query text existed was in links from other web sites.)
Link profiles are among the most influential ranking factors measured by search
engines. In competitive search query spaces, the web sites with the best link profiles
are most often the top ranking web sites.
Natural Link Acquisition
Link acquisition is vital to effective search engine optimization.
The rate of link acquisition is an important part of link building. Search engines
abhor manipulation. An unnatural pattern of link acquisition will appear manipulative.
Falling link acquisition can signal decline.
- Natural Link Growth – A steady rising and falling pattern, like a monthly
- Natural Link Growth – A steady increase consistent with rising popularity
- Natural Link Growth – A mostly steady rate of link acquisition (blue)
- Unnatural Link Growth – A one-time spike (red)
- Negative Link Growth – A steadily falling rate of link acquisition consistent
with falling popularity (brown)
It is also probable that each additional inlink has less influence. You could
say the 10th external inlink is worth more than the 100th
and that the 100th is worth more than the 1,000th.
External Link Distribution
The distribution of links to a domain is important too.
- On most web sites, nearly all the external links point to the home page or
to a small number of pages.
- On better web sites there will be lots of links to lots of different pages.
This signals to search engines that the domain has lots of different respected content.
This is one reason why blogging can be an effective content and link building
Search engines look for many signs of natural linking or for manipulation. If
your web site is found linking to other web sites unnaturally or benefiting from
manipulative links, the search engines may penalize or ban your domain. Some unnatural
- Reciprocal links – When two pages or domains link to each other.
- Site wide links – When every page on a web site links to another domain or
to one page on another domain.
- Hidden links – Links that are not visible to people because they have no anchor
text, are displayed in the same color as the background or made invisible with CSS.
- Paid links – Links that are purchased. (This does not include legitimate advertising
or directory listings)
Context can play an important role in identifying unnatural links. For example,
many blogs that cover the same subject areas and break new stories will frequently
cite and link to each other. This is known as echoing. Search engines do not consider
this reciprocal linking. In fact, if enough web sites link to each other frequently
and other web sites link to them, the search engines can recognize these sites as
a ball or circle of trust and give them dramatically increased authority.
External NoFollow Links
Search engines use links to measure trust, but their engineers do realize that
you may need to link to a web site that you do not trust or have not been able to
review. So you do not pass trust from your web site to these untrusted web
sites search engines ask you to add the NoFollow attribute.
<a href=”www.domain.com” rel=”nofollow” >Anchor Text</a>
Search engines also ask that you use the NoFollow attribute on links for which
you receive compensation.
Internal NoFollow Links
The NoFollow attribute can be used on internal links too in order to control
the flow of authority passed by links. If you web site home page has x authority,
and if it links to n pages, then each link passes (x/n)*d where
d is a dampening factor. Dampening is necessary so authority cannot be passed
forever and ever without end.
In theory you may want to NoFollow links to pages with little content. One example
might be a contact page that only has a form.
You may also want to NoFollow links in your page templates that link deeper than
the next level of pages down or to too far across into other topics or categories.
This is consistent with maintaining good silo or pyramid style web site architecture.
Using NoFollow links internally is a relatively young practice. Some terrific
debate exists about its efficacy and how best to implement the tactic. For example,
a recent question asked if you NoFollow a menu link with generic anchor text and
have a DoFollow link in the footer with contextually rich anchor text, will a search
engine spider follow and index the second link? SEO theory experts and investigators
are exploring these questions right now.
Measuring External Links
External links cannot be easily measured. While many SEO and search marketing
companies have attempted to quantify link profiles into a simple and meaningful
scale, it can actually be more accurate for an experienced SEO expert to look at
the quantity and sources of links and to use his or her judgment.
Measuring Link Numbers
The only way to count the number of links pointing to a web page or domain is
to crawl or spider the entire Internet. Only the major search engines and large
technology firms have the resources to attempt this. For this reason we rely on
data given to us by the search engines.
Many SEOs consider Yahoo! Site
Explorer to be the best source for external links. It will give you a count
of all external links and a list of up to 1,000 links. You can also download a tab
delimited file of up to 1,000 external links.
The problem with Yahoo is that, if a web site has more than 1,000 links, many
of the links that Yahoo reports can be from the same domain, especially if a domain
receives site wide links.
There is also a question about how accurate Yahoo! Site Explorer’s numbers are.
It can be argued that these two queries should show the same number of links, but
they do not.
Google Webmaster Tools is a great source of external link information, but only
for your own web sites. While you can get a terrific list of your own links, you
cannot get information about competitors. This makes it useless for comparative
Google regular search interface does have a link: operator, but it has been disabled
for many years. For example,
link:www.looneymaiden.com displays only three links. Even though Google’s link
counts still appear in many tools the data is worthless.
- Other web sites, such as All the Web,
provide external link data, though their search indexes are usually provided by
other search engines. For example,
All the Web uses Yahoo’s index.
Other sources seem to excel at finding made for advertising web sites.
Measuring Link Quality
How good is a link? Search Engine Optimizers have asked this question since Google
created the links economy by marrying
PageRank with its search
engine. Most SEOs will agree that they would rather spend the same time attracting
a few high quality links than large numbers of low quality links.
Unfortunately for us there are few signals of quality we can measure directly.
For this reason we must rely on third parties. Here is a short list of examples.
- Google PageRank – Measures
the likelihood that somebody will find a specific web page by hopping from page
to page and only clicking on links. That definition is certainly inaccurate, but
I think it captures the spirit of PageRank, so it will suffice as an informal descriptor.
- Yahoo! Site Explorer Page
Links – How many pages link to the page that links to yours? You can also measure
things link links to the entire domain and how deep a page is within its web site.
Compete are three web sites that try to rank
every web site from most visited to least visited. Supposedly, a link from a web
site on the top of their lists will be more valuable than a link from web sites
low on their lists. None of these services is wholly accurate and you should take
their exact results with skepticism. That said, if all three trend a web site highly
or as growing in popularity then it is probably true.
- Links or citations in sites like del.icio.us
or Digg or
Technorati can indicate popularity.
- Domain Age – Search engines tend to like age factors because they cannot be
- Does a web site have links from government or university or other types of
web sites that search engines generally trust highly?
The list could and does go on and on. Any signal, factor or variable you can
capture and measure is fair game. But what happens when a web site has 100,000 links
or some other huge quantity? How can someone possibly assess every link? The answer
is that you cannot. Which is why experience, intuition and gut feelings can be so
important. It is one of the reasons I call SEO a craft, part science, part art and
a lot of honing.
It’s getting late and this post has gotten long, so I’ll stop here. If I have
created more questions than I answered, if I have confused you, or
if you want to know more, ask away and I’ll try to answer your questions.