Google sitelinks have been a hotly debated topic over the last year and there are many theories as to how they are generated. However as a result of numerous queries from clients I have delved into this deeper and posted my findings below broken down into two distinct categories facts and theories.
THE FACTS
1. The facts are fairly simple. Sitelinks are a list of links shown below a search result. There can be up to 8 links below any one result. I have pasted an example here:

2. Google produces sitelinks by an automated algorithm. Their intention is to provide users with a shortcut that can possibly save them time when searching.
3. Google analyses the link structure of a site to find good sitelinks. Some sites that do not have a clear structure may actually prevent Google from producing sitelinks. This process is completely automated.
4. You can block sitelinks from within your webmaster account. This means if Google’s algorithm has got it wrong you can remove the short cut.
THE THEORIES
So that now leads me to the theories. Most of these surround the mysterious algorithm. However it is fairly easy to test some of these so I have had a go and would really like to hear from anyone that can prove or disprove any of these:
1. You must rank first in the natural listings for a term. This may actually be a fact as I can not find any examples to disprove this.
2. Traffic has to be of a certain level. I have checked all my clients that have sitelinks and found the lowest monthly traffic volume to be around 13,500. Interestingly all the clients that have sitelinks with traffic volumes less than 20,000 have very high traffic volumes for their branded names (Usually above 2000 per month). However clients with over 20,000 visitors per month can get away with having less than 500 searches for their branded name. This may just be a coincidence but could be indicative of a dual threshold for overall traffic and traffic for a particular term. It could be that traffic above 20,000 per month triggers sitelinks as does a unique term with over 2000 searches per month. I am very keen to hear about examples that would confirm or disprove this.
3. The click through rate will have to be very high. I have assumed that this would have to be above 90% but have no data to support this. My reasoning behind this is that I have only actually seen the sitelinks for terms that relate to a website URL or a brand name. I am very keen for anyone to post examples of any sitelinks that are generated for terms that do not relate to brand names.
4. The age of a domain seems to be an issue. I read in one blog post that the domain had to be several years old. However we had a client recently achieve sitelinks whilst the domain was less than 2 years old. I also have another client with traffic over 15,000 and brand name searches over 2000, however the site is less than 1 year old and as such does not have any sitelinks. As such we can deduce that the domain has to be at least 1 year old but can be less than 2 years old.
5. Link popularity of the site and deeper pages must also be a factor. This would help Google decide which pages were most important.
6. The structure of the site itself is very important and if not quite right will actually prevent sitelinks from being generated (This much we know is a fact). However from my research the sites that gain sitelinks seem to have a clear structure, a templated design (with the template present on the home page) and more often than not a breadcrumb trail.
So whats all the fuss about?
For a site owner sitelinks may be seen as an accolade that has been earned. It is almost a compliment or a statement that Google has taken note of your website.
At the moment the only examples I can find for sitelinks are for branded terms. And I think this is a crucial point to note. Could sitelinks serve another purpose other than helping users find the information they want more easily?
The algorithm seems to be very good at recognising unique brand names and attribute them to your site. Would this algorithm with a bit of tweaking be able to specifically identify unique brand terms and use this as a factor for page rank? The mention of a unique brand term in a page could be measured in a similar way to backlinks? This would add an extra dimension to search marketing and mean that branding would become much more central to all search marketing campaigns.
Finally make sure you check your site links
If a your site does receive sitelinks it is well worth checking exactly what Google is predicting your clients are after. Have a look at this search result for cineworld.
There are definitely some that that need to be removed here and in fact these links could actually be adversely effecting user experience. However there still is no way of telling Google what your preferred sitelinks are, this may be something that is just around the corner.
4 Responses to Google Sitelinks: Facts and Theories