Research & Rankings

Hi all,

I'm starting a new site to do a few SEO experiments to help try and draw a few more logical conclusions for myself on a variety of on-page seo factors... Anyhow, in doing so I'm going to use very obscure keyword terms that only my site would have on it (i.e. made-up terms that currently yield zero results in Google). I will then attempt to analyze the effects of tweaking a number of factors, by creating several identical pages that merely have a slight deviation in one on-page factor.

My question: Do we know for certain that Google ranks pages by relevance in an intra-site search? I.E. if I search for "crazy search term" on site:www.domainname.com and I get 10 resulting interior pages, can I be certain that the first resulting page is what Google believes to be the most relevant?

If not, than I'll have to rethink the way I evaluate the tests.

 

"crazy search term" on site:www.domainname.com

E.g. site:www.domainname.com "crazy search term"

Then yes, pages will be sorted in relevance for "crazy search term".

 

Thanks Shawn, but I think I didn't do a very good job of explaining myself. What I would is optimize let's say 5 pages for "skinny pink elephants from france" or something obscure like that. Then I would do the following search:

"skinny pink elephants from france" site:www.domainname.com

Which is where I would expect to get all pages from my www.domainname.com site that have the term "skinny pink elephants from france". Typically Google would show me only a couple but if I click the "repeat this search with the omitted results included" I would then see all the pages on my site with this keyphrase. Can I then conclude that Google sees the resulting page that is in the first slot as the most relevant of those pages within my site?

Another way to look at it might even be to take a real example... Lets say this site and the term search engine optimization. Does this search:

search engine optimization site:www.digitalpoint.com

http://www.google.com/search?source...engine+optimization+site:www.digitalpoint.com

... does that not only yield resulting pages but does it also tell us that the DP tools page is in Google's opinion the most relevant/best-optimized/best page on the entire DP site for the search term - search engine optimization?

 

Great thanks, I was replying to Shawn's post as you were posting this...

Thus I could have 5 pages all optimized for "crazy search term" and all be essentially identical (and very basic overall) pages and I could change one thing on each page and evaluate Google's preferences for various on-page factors based on this? Meaning that I could have 2 pages the exact same and on one page I could make the first instance of the keyword BOLD and keep the second page as it is and upon indexing draw a very simple conclusion based upon whichever ranks higher?

My real purpose would be to evaluate things such as keyword density and cleanliness of HTML code, etc. on rankings. In practice I might try pages with varying KW densities to try and determine at what point higher KW density stops "helping" rankings and starts to "hurt" rankings by setting off a filter or penalty, etc. Naturally its much more complex than I've described, but I just want to make sure a site:www.domainname.com "crazy search term" search is a reliable method for comparing Google's preference of pages within one website.

 

Exactly. Try to be careful with your links tho, make sure linking is very consistent and fair.

 

Yeah definitely - I'm going to use obscure terms and only evaluate pages within my site so that external links don't enter the equation. I plan to just have outside sites link to the homepage and from there I will link to the internal pages that are the testing pages and I'll make sure to limit variations in internal linking. Thanks all!

 

I think there's a good chance that the dup content filter will distort your findings.

Also, I suspect that there are many different routines that analyze many different attributes (e.g., density, prominence, titles, IBLs, OBLs, etc.) that each run on different schedules making it extremely difficult to draw any conclusion from changes made.

My guess is that 5 different sites (IPs) with varying content but hitting the various densities you want to hit, would yield the most accurate results. I’d be prepared for a long experiment. In my minimal experience, I’d argue that it can takes months to see the effect of a particular strategy.

Then of course, there’s the dreaded sandbox. ugh….

 

In relation to the Duplicate filter. After researching a little bit the questions related to the Dec 31 1969 cache dates. IMHO I believe that these dates are a sign of some type of duplication.


Looking at a competitors website I viewed all 355 pages indexed by Google. I found several pages that had the cache dates of Dec 31 1969. When I searched for the terms related to those pages I found that there were pages showing up in the results with the url strings associated to the 1969 cache pages. Here is the Newer page Indexed and here is Older Page Cache

It seems that google is indexing the new pages and not the old, however the old pages are the ones showing up in the search results with correct cache dates on them. From what I can gather it seems that the age of a page is a major factor in the results.

As I had stated on another post all my Blog pages have a cache date of Dec 31 1969 and after checking some of those pages against digitalpoint and a few other blog feeds where my Blog link is displayed, it seems DP and the other blog feeds are outranking those new pages and they are indexed by Google for my website. So this leads me to believe that the age of the website or the page showing this content, the older one will rank in the results until such time as Google removes the Dec 31 1969 cache filter.

I will also state I do not see anything that would point to link popularity as being a factor from the example I gave above only the age of the page.