http://www.webmaster-talk.com/showthread.php?t=24763

The above thread has suggested an experiment into testing Googles SERPs algorithm. I'd like to see a show of hands of all who'd be interested in participating.

The goal: Measure the importance of one-way incomming hotlink text for 2 scenarios.
a. A SEO'd page for a highly competitive key phrase (currently "web design").
b. A webpage that does not feature the key phrase (currently "ipsum") in the source or content at all.

Method: A yet to be determined site owner/s uploads 2 pages for the experiment. One that fits scenarios a & b mentioned above. Board members participating in the experiment then link the pages using the key phrase in the link text.

Analysis: SERPs are monitored by forum members and findings discussed to make solid conclusions based on stats. The most effective stat breakdown to monitor has yet to be agreed upon.

Requirements: Webmasters who host the experimental pages must have an established site which is not currently sandboxed and is regularly spidered by google. At least 150 participants are required to hotlink the pages to make the experiment viable.

Webmaster must not stand to gain commercially from the experiment.

Please post your feedback, this is an flexible structure and if you have an idea that will improve the effectiveness of the experiment please suggest it.

We all stand to gain from this action. The results of which will also prove to be a traffic puller to webmaster-talk.com once the results propagate throughout the web

Well, obviously, since I did propose this idea in the first place in that thread you linked to, I'm in.

I've got a network of sites that I'd be willing to post links on, and a plethora of client sites that I'm going to request do the same.

However, all of my sites are commercial (sad, isn't it?), so I don't think I'd be able to be the one who actually gets linked to. I did recently buy the domain name webdesignarticles.com and I haven't yet done anything with it; maybe we could develop this into a simple one-page site for the purposes of this experiment? Web-related keywords are the among the most competitive out there, so it should be a challenge for us.

Alternatively, I could register a domain name with a random string of characters as the URL and fill it with content that is completely unrelated to the anchor text that is going to link to it. This way, the only factor involved in its rankings will be the inbound anchor text.

I know this isn't exactly what you had in mind ("...must have an established site... not currently sandboxed and is regularly spidered by google..."), but it would make it more of an official experiment by reducing possible lurking variables.

I'm going to post a link to this thread in my signature to help advertise it; I recommend more of you do the same!

Update on my previous reply: that was just an alternative suggestion; if someone out there has a completely ad-free non-commercial site that they would like to use as the dummy in this experiment, that could work too. However, if it gets decided that starting a brand new test site is the way to go, I'll cover all domain registration and hosting costs myself.

I'm in. Sounds like an interesting idea.

The hosts of the web pages wouldn't necessarily have to be non- commercial, the targeted key-phrase could be non-commercial and non related to hosts' product-services. It's just important for the host site to be an established domain which is regularly spidered by google in order for the experiment to work in a timely manner.

Alright, fair enough. Do you by any chance own any non-commercial domains?

I'm brand new here, but you are welcome to conduct this trial on one of our charity sites. All are out of the sandbox. The one I would put forward is: www.homeless.org.au which is top ten on most data centres for a 9 million result search for 'homeless' so well and truely out of the sandbox.

Otherwise, just let me know where it ends up and I'll donate a few links from our sites towards the trial for it's duration.

http://www.webmaster-talk.com/showpo...58&postcount=8

I have posted a reply there.

Alright, Dominic, if you're up for it that's good. Your site should work fine; I suppose the next step is coming up with a keyphrase to target.

As crazy as this might sound, how about we target "web design" as our keyphrase? We all know how hard we've tried to get to the top for a phrase like this... let's see if we can do it with a site that has nothing to do with design at all!

shabbirbhimani, would you also be interested in this experiment?

Originally Posted by DevelopedWeb

shabbirbhimani, would you also be interested in this experiment?

Yes I am

If the target phrase is "web design", the effects will be hard to measure (25000000 results). There are 1000's of sites that have agressively targeted that phrase for years. I believe that phrase is out of the scope of this experiment. Perhaps we could expand the scope of Goal A to monitoring how many links it takes to get the page to google's top 10?

For experiment page B, my proposed key term is "ipsum" (623000 results) at google. Post a 500 word content page regarding the weather in Florida or something. Perhaps webmaster-talk.com might be interested in giving us a weather page for the duration?

p.s. I have a conflict in interest in hosting either page A or B. Any site will stand to gain at least exposure, which in any case is usually good. All my sites are commercial in nature.

What Data
So what kind of data do you want to track?

Inputs:
1. URLs of pages that link to each target page (including commencement date of when the links were added).
2. Toolbar PageRank of these urls

Outcomes:
1. Rank in search results for each page (google, msn and yahoo)
2. Allinanchor rank for each page
3. Resulting TB pagerank of recipient pages

I suggest we pick 1 data centre from google (up to 3 if it's believed to be crucial), and also follow the serps in yahoo and msn.

Reporting
I can publish a timeline of progress (inputs and outcomes), a complete report on another page on the site and a summary in this thread (if that's what you want).

For the sake of everyone's sanity, can we limmit reporting to three days a week so it's not something I need to produce daily?

Top 50 Sites
I'd also like to watch the movement of the top 50 sites in the serps for the duration of the experiment. For this, I'm suggesting we plug all the urls into the digital point keyword tracker and let it do the work, and I'll just publish the chart it produces in the report.

Someone who is willing to share the workload may want to volunteer to input the urls into the tool and check this daily. We could then include the graph in the report three times a week (any takers)?

Web Design
I believe 'web design' is a good choice as I'm only really interested in how this works out with a compeditive phrase - don't know about everyone else.

Writer Needed
Who is going to create the two pages?

Link Building
Where are these links from? Only a few people have participated in this thread - is anyone willing to 'put the word on the street'?

Metho, we posted at the same time, so I missed your post.

Ipsum is fine with me, but will a search with only 7,000 allinanchor results be adequate?

Since you mention it though, any search term should do and we would be able to see the results faster with ipsum.

BTW, where the **** did you come up with that!

Dominic, I like what you said about using a more competitive phrase that "ipsum." What I would like to do is try something new; others have succesfully targeted non-competitive phrases such as "miserable failure," perhaps it's time to try this experiment with a more competitive phrase.

Originally Posted by dominic

Writer Needed
Who is going to create the two pages?

I'd be willing to put together the content (it doesn't even matter what the content is for the sites anyway, so this should be the least of our problems). Could you please explain to me what the need is for having two content pages now as opposed to one?

Originally Posted by Dominic

Link Building
Where are these links from? Only a few people have participated in this thread - is anyone willing to 'put the word on the street'?

If you have a blog, this is what it's good for. Also, as I said in another reply, I'd recommend that anyone who participates in this puts a link in their signature similar to mine. Once we get a little more support I'll PM the mods about perhaps making this a featured thread.

Originally Posted by DevelopedWeb

Could you please explain to me what the need is for having two content pages now as opposed to one?

I took from this:

Originally Posted by metho

The goal: Measure the importance of one-way incomming hotlink text for 2 scenarios.
a. A SEO'd page for a highly competitive key phrase
b. A webpage that does not feature the key phrase in the source or content at all.

That two pages would be made, both linked to in exactly the same way (one with on page optimization and use of the keyword, one without optimization or use of the keyword).

I assume them being on the same domain is fine. If it needs to be on a second domain I have another non-profit site (www.life-coach.org.au) that is top ten in a 7 million result search, also well out of the sandbox.

The Phrase
I too would like to stick with 'web design.' We don't need to hit the top ten for web design - to draw conclusions from the experiment. Those serps provide the ultimate testing ground - a set of search results on par with or more compeditive than, the serps people learning from the experiment will want to apply what they have learnt from the experiment.

Hmm, to clarify. I was thinking of 2 key-terms/phrases for each page. If everyone's fine with "web design" for A, no worries. It just sounds like a lot of work to drill down 50-1000 pages of results to locate A's position.

I was hopeing "Ipsum" for B. We'll get immediate, incremental results and #1 is a likely result from page B.

*edit, btw dominic, lorem ipsum is a passage of latin text that has been used 'for placement purpose only' for centuries. Designers use this passage as dummy text in page designs.

metho, as nice as it would be to see ourselves easily achieve a number one position with something such as "ipsum" (or "lipsum", as others call it), that's not an original idea--others have succeeded with "miserable failure," "french military victories," etc. What would be great is if we could organize something new by using a very competitive keyword in our anchor text.

A question for you guys and anyone else who may be reading this: are you members of any other web-related forums? If so, get to posting! Rally support for this--are we all decided that we're going to use dominic's non-profit site as our "webpage that does not feature the key phrase in the source or content at all?"

The ipsum term for page b is the control group. All valid research methods must have a control group to benchmark results. The google algorithm changes constantly, so this experiment requires a fresh instance of similar, past efforts if page A is going to have a control set of results to compare to.

Metho, so are you saying we will be asking for 4 links from ppl?
p1 phrase 1
p1 phrase 2
p2 phrase 1
p2 phrase 2

If the goal is to compare the effects of on page factors plus ancor text v. anchor text alone, why can't we keep it simple and work with two pages on the one site, one with on page keyphrase and optimization and one page with no on page factors at all?
p1 phrase 1
p2 phrase 1

I don't believe you need a parralel control duplication of this trial unless you want to submit the results to a journal.