Misreading of #3 and Proposed Experiment

Thank you to the numerous individuals who pointed out my misreading of #3 in Rand’s list of SEO questions unanswered. The real question he was proposing, which I misread, was whether or not link removal could actually injure a site beyond where it would have been if the link had never been posted. While my initial reaction is no, I think a simple experiment is in order.

Proposed Experiment
1. Two pages (“Receiving Site A” and “Receiving Site B”) developed using the same 250 nonsensical 7 letter words with no title or meta on a new domain. (“Receiving Site B” will receive the treatment of losing a link)
2. “Linking Page A” will be chosen, with a link to both sites using a 8 letter nonsensical word.
3. “Linking Page B” will be chosen, with a link to just one of the sites using the same 8 letter nonsensical word.
4. Wait until the cache of both A and B are updated, both test pages indexed, and rankings stabilized for 7 days across datacenters.
5. Pull the link from “Page B” which points to “Receiving Site B”
6. Watch what happens.

#Note: Because we know that order of links on the page matters for the amount of weight passed to it, I will place the link to the experimental page (the one which will lose the inbound link) FIRST on “Linking Page A”, essentially placing a greater burden in favor of the null hypothesis. Otherwise, we risk a switch in the rankings of Receiving Site A and Receiving Site B because of the confounding factor of link order on Linking Page A, rather than a penalty for losing a valuable link

Expectations / Hypothesis:
I cannot rationalize a penalty for link loss. The strength of link order alone on Linking Site A will far out weigh any possible penalty for link loss, that no changes in rank order will occur.

Let the games begin!

No tags for this post.



  1. Technical SEO Explosion: Tips Pulled from SEO Experiments | Blog Posts - [...] (note - Russ didn’t answer the question as it was asked and proposed a follow up test) [...]

Submit a Comment

Your email address will not be published. Required fields are marked *