Reciprocal Linking Test

Hi

I am going to run a test on reciprocal linking. Here are my thoughts on how the tests should be conducted - please make suggestions on how the test can be improved and made as fair as possible. I want to be scientific about this.

Thanks

Test:

This is designed to test the impact of using recirpocal links
in conjunction with and as opposed to one way links.
This test should help to determine the importance of reciprocal linking
and answer a heavily debated topic with definitive evidence.
It should also go some way to helping identify recent algorithm changes.
The URLs will not be released until after the test has finished.
Anticipated length of the test is 2 months but it may require longer considering
the Google sandbox as the sites have only recently been submitted to Google.co.uk.
All of the sites will be optimised for the same non comercial made up keyword that returnes
a few results. This should make analysis easier.

The sites will be linked to as follows

Site 1: 1 way links only
Site 2: 1 way links & Relvent reciprocal links
Site 3: 1 way links & any reciprocal links
Site 4: reciprocal links only

Conditions:

All sites registered on the same day

All sites submited to Google.co.uk on the same day

All sites will have the same number of 1 way links from the same sites (except site 4)

All sites on the same type of hosting package on the same server

Update:

Sites 1,2,3 all have the same 2 one way back links using the targeted keyword as anchor text.

Site 1 has been included in the index today. It has taken a while to get the site indexed and still only one site has been included.
I am monitoring the indexing status among many other factors.

All sites were linked to at the same time from the same sites, all sites were registered at the same time and submited at the same time. They all have the same code.

I am wondering if a) its because site 1) comes up first and maybee I should mix the order in which they are linked and or b) that google randomly selects which sites to include in the index from its index rather than doing them in a sequential mode.

In saying that the volume of indexing would mean that other sites coule be included in between these sites even if theyw ere indexed together.

The links are from well established domains on different IP addresses.

Good work. Will be very interested to see the results.

Please do post some results. I am anxious to see what you will come up with ...

Update

As of yesterday site 1 & site 3 are indexed on data center # 72.14.203.104. Today the data center has changed to 66.249.93.104.

Site 1 was showing its self as a back link where as site 3 wasn't. I find this a bit strange.
In the new data center site 1 has all 9 of its pages indexed where as it only had 1 before. Site 3 only has one page indexed. Both sites are showing no link in google but they both have 2 one way links.

Interestingly site 3 (newly indexed) ranks #1 for the targeted term and site 1 is no where to be seen! ONe of the sites linking to it is also up at position #4.
The term is non competetive but does contain the word seo which makes me think it is being sandboxed because of that word especialy.

I think site 1 is now sandboxed and site 3 is experiencing an initial boost. I am recording the positions along with other statistics and also keeping a sandbox duration record. I can't wait for all the sites to get indexed so I can start testing linking out.

you should avoid making short-term conclusions - quarterly is likely best. The problem with a real-time, real world environment is - you never get to view all the processes involved, which ones are factored in which ones are not, what site caught a crawl through a link at the beginning of a cycle - what ones didn't, what ones are being pulled from 'fresh' and so many more.

true...you've got to analyse it during a long period and then, only then, you should start thinking about real conclusions...

still, it's an interesting experiment that i'll gladly keep on hearing about

Fathom, I appreciate your comments. I haven't made any conclusions , just observations. I am keeping notes as to what has happened and what my thoughts are on each area so that I can look back at the end when I do make conclusions.

It might be completely wrong but this is what I am thinking at the moment.

I will wait until I have all the links in place for a while and that all sites are on a level playing field before I make any kind of conclusion.

I think its important to make these observations as time goes on as it won't be as easy to remember everything that has happened at the end of the test.

nah - it's always good to toss ideas in for what you are seeing - to gather a volumes of opinions - just noting that many tests can be skrewed simply because the search engine didn't play as expected.

A few years back 'sig links' where challenged in numerous tests for PR passing, weight, and relevance... the tests themselves produced no desirable effects and conclusions 'suggested' nothing of any value occurred via a sig link.

The tests themselves though were flawed and months later it was shown "why".

In any case my professional opinion on yours - goes as this:

The 'link version' [one way/recip] matters less than how it is used.

A little 'encryptic' - so in plain english - a continuous pattern repeated over and over again - is worth less than changing it.

All links tend to go to the page "at the top" regardless of whether it is a "one way or a recip"...

The idea of "link bait" moves that pattern to other locations - as "link bait" rarely sits on the top page... and naturally gain 'one ways' however, recips can actually become far more powerful [than the simple 'one ways at the top'] simply because you changed the pattern and got the flow of "everything moving in reverse".

NOTE: there is nothing stopping anyone from asking for a link [to "link bait"] in exchange for a link back - course the pattern doesn't resemble the traditional 'reciprocal exchange'.

It is also worth noting that "top pages" are rarely "useful pages" they are merely the access routes to more important/useful pages deeper in the website [and on diversity of topics] - and thereby 'default' the traditionally placed link - it not likely the best place for a link.

tstolber,
I'm new to this, so forgive my ignorance. How do you know when a site has been indexed and what datacenter it has been indexed on?
Steve

Google obviously don't want their search algorithm reverse engineered. If they're as smart as they claim then they will have included a small random element in amongst the supposed 100+ ranking factors. On a normal search it wouldn't make much difference. However, on a controlled test such as this, where most of these factors are constrained to be identical then the random element could significantly skew the results. If that's the case then you'd have to run the test many times over to average out the randomness.

I CAN HELP YOU GET RELEVANT RECIPROCAL LINKS FOR SITE4 TO PROVE THAT RELEVANT RECIPROCAL LINKS ARE GOOD AND COST EFFECTIVE. I WOULD RECOMMEND TO CHANGE SITE 4 : RELEVANT RECIPROCAL LINKS RATHER THAN RECIPROCAL LINKS

Make sure that you keep us to up to date on your finding.

quna

Thats not a bad idea and its not too later for me to do this.

The difference between sites 2 & 3 should show how effective the relevant and non relevant reciprocal links are.

sgbphotography

When a site has been spidered you will see the googlebot user agent in the log files, this means the site has been indexed. When it has been cached and included in the index (slightly confusing read as database for search results) you will see it appear in the google search if you type the domain name. Put your mouse over or click on the cached page (link below each title and description in the results) and you will see the URL is an IP address. This is the datacenter that the results are coming from. There are many different active data centers.

h1926ch

I am keeping detailed records. I will keep everyone updated when anything significant happens.

Thanks for your interest.