If you're out there doing weird artificial stuff with inbound links (like buying them, or link farms) then I agree -- you should worry about having a "natural" link distribution.
However, the problem is, you and I have no idea what a "natural" link distribution looks like.
When I used to run Talkr.com, people would often link to the site with the link-text "Talkr" or "Talkr.com".
When I linked to the SearchEngineWatch a couple of paragraphs ago, it was much more natural for the anchor text to say "article" than "anchor text distribution article".
For the last few great feature articles that I've published, I've gotten a few links from very popular high-pagerank sites that made the editorial decision that it was a good story. But no links from little sites.
So what does a "natural" link distribution look like? You and I don't know. And I doubt that Google knows.
How would Google build up a corpus of websites that only had editorially chosen links, against which they could train their algorithm? (Perhaps it would be easier to build a corpus of websites that had spammy inbound links -- but the problem there is that good sides with spammy inbound links would also attract natural links, because they would tend to rise in the search results.)
Probably, there's a distribution of distributions. And it may be quite possible for Google to spot dramatically-manufactured link distributions.
But the irony is that there's a good chance that manually trying to tweak your anchor text or the PR of your inbound links results in easy-to-spot unnatural distributions like:
- Buying links based on pagerank formulas (ie buy x PR 5 links, 2x PR4 links, 3x PR3 links, etc.)
- Systematically varied anchor text (ie 20% says "mortgage leads", 20% says "mortgage quotes", 20% says "mortgages", 20% says "buy mortage", 20% says "insurance.com")
What do you think? Am I an idiot, or is this spot-on?
No comments:
Post a Comment