One time (at band camp) I manipulated first link priority to the home page of a site for the site’s main keyword – that is, instead of using ‘home‘ to link to my home page, I linked to the home page with “insert keyword“). Soon afterwords the site dropped in rankings for it’s main term from a pretty stable no6 to about page 3 and I couldn’t really work out exactly any other issue.
Of course, it’s impossible to isolate if making this change was the reason for the drop, but lets just say after that I thought twice about doing this sort of seo ‘trick‘ in future on established sites (even though some of my other sites seemed to rank no problem with this technique).
Confused as to the best practice?
So was I.
So I formulated a little experiment to see if anchor text links had any impact on an established home page (in as much a controlled manner as possible).
*****Setup – I EDITED THIS A BIT: Basically the search term (anchor text) I was looking for was “‘keyword1 (not present on the target page) + ‘keyword2 ‘” -(minus) ‘keyword3 also not on target page but a common word that would accompany keyword1 ‘” – I did this to get rid of a lot of noise in the serps for more relevant pages to the original keyword phrase. Also this keyword (keyword1) appeared in anchor text on only ONE internal link on a static page which had no other links to the home page).*****
Result:
Well look at the graph below.
It did seem to have an impact.
However that massive drop for months is kind of worrying.
From Jan to July the site was nowhere for the phrase. Although it has just jumped 105 places back to no3 for the test term (which was a geographic location – not a made up word).
This change could be down to other reasons as I said – Google is always tweaking things. Perhaps this ranking drop would not have happened if the keyword was present on the target page.
It’s very posible linking to your home page with keyword rich anchor text links (and that link being the ONLY link to the home page on that page) can have some positive impact in your rankings, but it’s quite possible attempting this might damage your rankings too.
Trying to play with first link priority is for me, a bit too obvious and manipulative these days, so I don’t really bother much, unless with a brand new site, or if it looks natural, and even then not often, but these kind of results make me think twice about everything I do in seo.
I kind of stay away from the more manipulative onsite stuff these days.
I thought I’d share and let you decide on the potential risk or reward, or comment on…..
These days it is very dangerous to try anything at all. Even adding content to my e-commerce site… I’ve seen my rankings going up and down like crazy without any clues. (like in the stats) It seems when I add a ton of content to my sites, my rankings shift all over the place taking months to become stable again. It’s like google is trying very hard to determine what the site is about by reading links and content, but in my case I can only say: “that’s cute but it’s WRONG!” It seems that time is always a factor in everything you do, same in the experiment above….. Interesting thing though this first link item… I wonder if the first comment on a page (visible in the HMTL) does the same.
“‘keyword phrase‘ -(minus) ‘keyword I new was not on the page‘” I “knew” :D
Thanks Erica yeah my spelling is terrible on this blog! I edited that whole bit to make it more clear.
Lots of external links to the target page using the anchor text? This test tries to isolate too much and doesn’t take into account other factors happening on the domain (such as ranking for other more important stuff which can have a massive effect on other pages) Far better to make this change on a site that has hundreds of links for the term and is the site’s core focus (or one of them), maybe with a regional modifer. Then roll that out over 500 sites for a good sample set. Need to also be careful to monior what happens with all keywords on a site.
i know you like this sort of thing Andy. :) It’s very hard to isolate anything in seo as we both know…. I just wanted to see if it was possible if this sort of thing had ANY impact on the home page. It did… of sorts. It’s left me with questions too. PS – This was the same site I tried to manipulate with site-wide links to the home page that tanked.
Shaun, I love when you publish these tests! I think I understand what you did. You had ONE internal anchor link with a “special” word that didn’t show up on the target page, and NO other links to that page had this special word. The only way the target page would rank at all for that special word is if google “counted” that internal link. Now that said, I don’t understand the yellow-green versus green dots on your graph. And are you saying the result was HORRIBLE for 6 months and then it went back to what it was before (so no benefit)? Can you clarify what the two lines are on the graph. Maybe I’m missing something obvious. :) Steve
Yes Steve you got it :)
Google UK rankings, and Google UK rankings when you select “only from the UK” set of results…… The graph shows there DID seem to be a benefit (that’s when I started tracking the keyword you see…. when I saw it pop in immediately. The graph also shows the page DID NOT RANK after a short while, for many months, and then just recently popped back in. I thought I’d share. From this, it would seem Google does count internal links to the home page on some level. I did surmise in the post that perhaps the massive dip in rankings was because the target word was not on the target page, only in anchor text links pointing to it. I also mentioned in the post playing about with this sort of thing could in fact hurt your rankings, but again, it’s very hard to identify. I published it in the hope of generating some discussion on this element of optimisation – there’s not much in SEO (at a granular level) you can clearly identify has a boost in rankings. I get a lot of mixed signals no matter what way I set up the test. I have a few other ideas though to follow through on I will probably try out in my xmas break :)
Interesting. That might explain some things, I wonder about sub-pages as well as that is the most typical place I would do that kind of thing
Wow, This is a really scary test. I’ve made this change on several websites and never seen any negative results. Although the positive results have been non existent as well. However, when crafting new websites, I think I’ll continue to use anchor text for home links where reasonably possible.