Tue 26 Jun 2007
Does Google know I am manipulating things?
Posted by ShaunzoFrom my observations, if you have a few sites, interlinking them isn’t a bad thing. But what about a lot site-wide links between sites?
Getting suddenly hundreds or thousands of links to your site from one site in one go is rarely someone else’s doing. Even when you comment on someones blog and get site-wide links, this is in effect, Your Doing. Your Manipulation. I can see how Google could view this, and I can see how Google could devalue your site. The algo can easily take care of this. Google doesn’t like (some) manipulation.
I’ve heard reports of people’s site’s tanking in all the search engines because of this practice. Some recover - some don’t. Why?
Trust, Relevance, Authority.
If you’re site is trusted, or the site you’re getting the links from is trusted, site-wide links might be a-ok (emphasis on might, I’ve not yet managed to get the W3C to give me a site-wide link yet, but I would surely accept them!). Why? Simply if your site is “trusted” - an authoritative site on the subject - you can get away with a lot more.
I’m a great believer in themes, neighbourhood and relevance (regardless of No-follow) when it comes to getting links. If people are saying Google isn’t good at this yet, you can bet your arse they will be “tomorrow”.
I’m also a believer in Single IP Address + One Link Hitting (well, get a few on strong pages) 
so sitewide links for me aren’t that attractive anymore.
What’s the benefit of a sitewide link? err…I don’t think there is much, if any. I think a sitewide link could devalue the site they are on and in turn add no real benefit to the site it links to, unless of course there was a strong “theme” to be detected on both sites. Relevant links - well that might be another story altogether.
Note I am talking about site wide links (all identical anchor text). I’ve done a bit of experimenting with sitewide links with different anchor text, and the results I have to say weren’t bad. I’ve also experimented with links on 90% of the pages
I think sitewide links aren’t as good as links that feature on a % of the pages, or indeed links that are on the strongest pages of a site. Sitewide links also kind of have the same kind of footprint as “paid for” links, which Google likes as much as I like the painters in round at my girlfriends
In the end, you must ask “does Google think I am manipulating things?“. Are you signalling to Google what you want to rank for when, in short, your site doesn’t have enough authority or trust to rank in the top 10 for that keyword yet? Cheek!
Sitewide links = visible manipulation, the only exception being, I think, two trusted sites in the mix, in the same neighbourhood, compelled to link to each other in this way.
On a final note it’s worth remembering if sitewide links are easy for Google to isolate and identify, they are easy to turn off or on when Google factors your site in it’s serps. Sometimes you’ll be up, sometimes you’ll be down, sometimes you’ll be out.
Do you really want your online marketing efforts so easily susceptible to such fleeting and easy to change nuances? I didn’t think so…..
So say no to site-wide links, unless they are from the W3C
June 26th, 2007 at 8:49 pm
[...] I used to be a PR monger - it’s a throwback to the good all days. A lot of people think PR is dead, some, including myself, don’t. As in the article I linked to says, Matt cutts says PR is good to get you out of Google’s “B” index, the “supplementals”. Sitewide links? Well I whittered on about those in this post about sitewides. [...]
June 26th, 2007 at 10:18 pm
Thanks for the advice. Up until recently, I pushed the sitewide scheme, but I can certainly see the pitfalls of carrying through with this practice.
July 10th, 2007 at 12:55 pm
What you should keep in mind with repetative link text is that you could have 20,000 incoming links with your comapnies name as the link text. Would Google stop you ranking for your only company name because of this? No. For this reason, I don’t think that Google will ever penalise highly repetative link text.
Google do ignore repetative link text and links to the same URL on one domain. I have a site with about 160,000 incoming links from a farily weak domain and all the links have the same link text, but the site doesn’t rank for that factor alone.
I don’t think Google currently does much about shared IPs. I know one site that has about 20,000 incoming links all from a handful of IPs but different names and they have been result one for their search term for a long time.
October 3rd, 2007 at 7:56 pm
Interesting point this. I know A LOT of sites that are no. 1 in the serps purely because they’ve built up a stack of site wide links. However, I agree that they are leaving themselves wide open. A quick algorithm change and it’ll be bye bye for sure.
November 4th, 2007 at 11:18 pm
I agree with the previous comments.
I have a fairly unique handle that I’ve used on the web for some time. On some sites the handle is part of the link text to one of my sites, on others it’s just in the page text. On the site in question there is not a single mention of the aforementioned handle but when searching Google for the handle the site is returned at position 1. I think link text and on page content is very important for ranking for your chosen keywords.
If the link is on a page that has nothing to do with your keywords then you can’t expect it to have much of a positive effect.
December 13th, 2007 at 1:03 am
I’ve been optimising for some time, and this is what I think.
Site wide links cant hurt you, but they may not do you any good.
If a site wide link on a poor site with 1,000 pages was bad for you it would be easy to trash your competitors website. put up a site with 1,000 pages with repeating content thats nothing to do with the theme of your site and link to your competitor. great idea instant competitor beater.
Do you really think Google would allow this obvious weakness in their algo? I dont think so.
If you have a site wide link its probably going to have reduced effect as the number of pages increases, up to a maximum value. So for arguments sake lets say 20 pages. If you get a sitewide link from a website with 10 pages then all 10 pages will count, but if a pattern is detected then the value of the tenth link will be less than the first. this way if you have a sitewide across 1,000 pages you get no value/effect after the first 20 links. This way Google can control the value of sitewide links and devalue them as they increase, but withot allowing someone to trash your site credibility. So they will not criticise/judge your website based upoon what someone else may do to you. But you will be HIGHLY criticised for your own content and your outbound link relevancy, because this is directly under your own control.
Ultimately the problem with lining theory, is its only theory, nobody really knows, BUT this does seem a sensible approach to me and there is a lot of evidence to back this theory up. Everything you have said about contextual relevancy (no matter how many pages) I agree with.
January 16th, 2008 at 6:40 pm
Good article, has actually spurred me into tidying up a few sitewides that one or two of my sites have - the rest of what you say on linking strategy is very good too, relevance is key.