Why there is no even close correlation between MajesticSEO data and Open Site? Explorer?

[eladlvy]

I compared my site home page against my competitor homepage with Open Site Explorer and I found that according Open Site Explorer I have only 3 backlinks where in MajesticSEO I have more than 100 backlinks in 'Historic' data and more 90 in '30 days fresh data'.

Why there is no even close correlation between MajesticSEO data and Open Site Explorer?

How SEOMoz count backlinks?

[mikecorso]

What's so awesome about a fresh index when it thoroughly confuses the issue (includes dupes from weird sessionIDs/URL parameters, etc). If MajesticSEO isn't showing accurate link count numbers, I have no use for it.

Mike Corso
Cool Site of the Day
The Internet's Oldest Directory
http://www.coolsiteoftheday.com

[randfish]

We did a bunch of work on this a while back, so my stats are probably not up to date (I think around October of 2010). Basically, we compared Yahoo! Site Explorer numbers, Google numbers (via the crappy but somewhat proportional link: command), Exalead, SEOmoz, Majestic and Alexa.

Majestic was definitely odd and so was Alexa. Neither of those two mapped/correlated well to the quantities reported by the others. Thus, for example, if xyz.com has:

• 50 links according to Google
• 1,000 links according to Yahoo!
• 500 links according to SEOmoz (Linkscape/OSE)

And site abc.com has:

• 100 links according to Google
• 2,000 links according to Yahoo!
• 1,000 links according to SEOmoz (Linkscape/OSE)

The percentages of quantity will match up fairly well for these, but not for Majestic (e.g. they might show 5,000 links for xyz.com and 3,500 for site abc.com).

This is a bit odd, but we don't know exactly why. They crawl a ton more links than even what Google/Yahoo!/Bing reportedly do, which could be part of it, but my best guess is the canonicalization and freshness issues. Since MJ crawls the web all the time, and doesn't build "indices" every X time period (like Google/Yahoo!/Bing/Linkscape), but rather maintains a single consistent link index to which new sites/links are added, the data structures may be different.

Majestic also appears, at least to us, to do far less canonicalization, removal of unnecessary URL parameters, etc. Thus, if one site links from 50,000 pages due to weird session IDs in the URL, that might bias the crawl and link count, but standard canonicalization will normalize these.

I really don't mean to bash on a competitor - MJ's fresh index is awesome, their tools are good, and a ton of SEO folks find them useful, including myself. But, on the index matching and link count numbers, we definitely see this same weirdness that many other SEOs do.

Hope that helps!

[qlkasdjfw]

Opensiteexplorer more or less takes a representative random sample of the available links. More prominent links have a tendency to be visible on opensiteexplorer rather than majestic seo. Also check out yahoo links.

[firstconversion]

Can you provide the site url?