09 Jun 2009

Pagerank is Dead. Sculpting is Dead. Trust is Alive….in my eyes.

There’s been a few debates that I’ve wanted to weigh in on… one that comes to mind is the big hub bub over pagerank sculpting via use of the nofollow tags….

Pagerank sculpting is based on the formula from the original Pagerank patent from 1998, (PDF) that says the value of a backlinks is found via dividing the number of links on page by the value of a page (with a slight dampening factor).

The public news of Google passing value a different way than this old formula on internal links can be found from the Aimclearblog coverage of Danny’s Q and A with Matt at SMX:

Matt: If you have 10 links and 5 are nofollow, there is this assumption that the other 5 get page rank. That might have been partially true at one time, but that’s less effective these days. You’re not going to get a penalty, or get in trouble, there’s better uses of your time. If your using Nofollow to channel page rank around your site, it’s like a band aid, focus on designing your site  purposely to sculpt Pagerank.

…Initially if you had 10 links, and 5 were no followed, the other 5 would get the remaining page rank, it’s not that way these days. It bubbled up from the indexing team and  it could change in the future.

….Q: So there’s 10 links on home page, 5 pages Nofollowed, where is that link juice going?

Matt: You can think of it as evaporating.

Q: It’s actually hurting your site then?

Matt: Use it sparingly, use it for links you can’t vouch for, if you are a power user and there’s a page you don’t want (sign-in) that’s a fine page to use Nofollow on. The only thing I Nofollow on my blog is a subscribe link, which is not all that useful for the main web index.

Q: If there’s a Nofollow page from that main page, there’s evaporation?

Danny: When page rank came out, 10 links on the page each got 1/1oth of the link juice.

Matt: Page rank has changed over the years, academic papers on this are fantastic. Our models and the way we compute it and the way we determine is more sophisticated than when the original papers came out.

I for one am not surprised that a formula dating back to Google’s first patent in 1998 is defunct… pagerank, as outlined in 1998, is dead…has been dead…sculpting something based on that forumula is dead…. the only thing that’s not dead about that forumula, is the visiable toolbar they show people  (and only SEO’s tend to look at that green bar).

I think the nofollow just got outta hand where people started nofollowing pages a bit differently than how google envisioned they’d be used for (first for “helping reduce blog spam”…then for paid links…then SEO’s started to use it for “Pagerank Sculpting”) …and Matt now comes in and basically says “stop wasting your time…. we don’t pass value the old ways anymore.”

……..

ok…so let’s play a Jim Quiz Game here….You’ve got 3 urls to choose from here.. and I can give you backlinks on only 2 of these 3 urls … which 2 pages would you like a link from?

Page A: (FAKE URL) http://www.bobs-blog.com/water-pressure-whoknew.html
Pagerank 5
Number of links on that page: 10 internal, and 10 going to other sites. These 10 sites are “low to so-so” on the Trust Scale.
Backlinks to that page:  100 internal links….. no links to that page from any other site.

Page B: (FAKE URL) http://www.nasa.gov/science/Science-of-water-pressure/resources.html (FAKE URL)
Pagerank 1
Number of links on the page = 100 – most of these links are to high trusted sites
Backlinks to that page ~15
(5 other nasa.gov pages, another .gov link, 2 edu’s, few others…those pages  show either 0 pagerank, or pagerank 1′s. )

Pager C: (FAKE URL)  http://library.harvard.edu/science/water/water-pressure-facts.html
Pagerank 0 – but has google cache
Number of links on the page = 100 – most of these links are to high trusted sites.
Backlinks to page 3 (2 other internal edu pages, and a dot com…none have pagerank, but all 3 are cached)

If you could only get links from 2 of the above sites, which ones would you take?

…..if it were me…..I’d take choice B and C over choice A.

Choice A would be the obvious choice using the traditional logic of (PDF) the original pagerank patent from 1998 (higher pagerank, and more value going to your link since there’s only 10 other links on that page)… but is that green bar the real value of the page?  (by value I mean Ranking Value)…and how is Ranking Value passed out on a page?…and Matt’s words echo in my head

Page rank has changed over the years, academic papers on this are fantastic. Our models and the way we compute it and the way we determine is more sophisticated than when the original papers came out.

So what do I theorize that Google uses for it’s model of finding the value, and passing of value in links is these days?

I believe that Trust plays a big role in Google’s model. I’m a believer in Trust Value, or what some call TrustRank (here’s a great video of Rand explaining TrustRank) (and even though TrustRank isn’t Google’s patent, I do believe Google is using many of those basic ideas).

In the TrustRank Paper(PDF) they talk about Humans identifying 200 sites (or parts of websites (for example, one of the 200 sites might be the library section of stanford.edu)) that are truly pages that link to lots of other sites, and are highly trusted links (pages that haven’t been manipulated by SEO’s). The more closely linked you are to the original trusted sources, the higher your trust score is going to be.

….but many people still look at manipulating pagerank… when I don’t believe that pagerank (based on the original formulas) is the ranking answer any more… I put much more faith in Trust Value than I do in Toolbar Pagerank… problem is, you can’t see Trust Value in a toolbar…  you can only see Pagerank 1998 in the Toolbar.

What if there were instead of a Pagerank bar, a Google Trust Value Bar… and let’s say that every page on http://library.stanford.edu is a TrustRank 10, and everything they link to is a TrustRank 9, and everything they link to is a TrustRank 8…..and if we could see that Trust Value flow…. then everyone would be running around saying, Damn… I need some high trust value links… where can I get some phat links from sites like k12′s or gov’s or edu’s…. but it’s good of google to keep those types of pages a low pagerank, so the brokers won’t come in and ruin the trust of those pages.

That’s partly how I think of it… a page may be a PageRank of 1 and it might have 100 backlinks on it, but it may be a Trust Value of 9, and every link might get the same trust level… no matter if there’s 2 links on that page or 200 links on that page.

…And since they’re no Trust Value Bar, I’ll use my ninja senses to value the trust of a page and a site when evaluating the value of obtaining a link…but I won’t get distracted by the Pagerank, nor the number of links on the page.