The Basic Principles of Peanut Butter SEO

by Michael Martinez on August 21, 2008

Peanut Butter SEO. I’ve used that expression quite often over the past few months because it encompasses a whole new generation of search engine optimization ideas. I have (rightly or wrongly) attributed the basic concept to Matt Cutts, who used a peanut butter metaphor at SMX Advanced 2007 to descuss PageRank.

Originally, all I reported was that Matt Cutts said you get “only so much PageRank for your site”. In fact, he went on to compare it to peanut butter, suggesting that you had to be careful not to spread the peanut butter too thinly.

Unfortunately, the SEO community soon after fell in love with the ill-conceived “PageRank sculpting” notion, using “rel=’nofollow’” to attempt to control the flow of PageRank. The whole debate over the use of nofollow on internal links distracted the SEO community from several very important messages that came out of Google.

First, Google has acknowledged that there is a limit to the amount of PageRank flowing around in its index.

Second, Google has admitted that its Supplemental Results Index doesn’t get the “full” treatment. That is, Supplemental Pages are not fully parsed and indexed.

Third, Matt Cutts revealed at Domain Roundtable 2008 that Google attempts to strip domains of their earned PageRank if they change hands.

Fourth, Matt also disputed the idea that simply adding more pages to a site will increase its PageRank.

Peanut Butter Search Engine Optimization has to focus on collecting PageRank because the more PageRank you accumulate, the more pages you can expect to rank well in search results (assuming you let the PageRank flow naturally through your Web site). That is, attempting to sculpt PageRank defeats the whole premise of Peanut Butter SEO.

Let me put it another way. You have pages that are in the Supplemental Results Index. How do you know which pages? They would be the ones that don’t appear in your site searches for expressions you know appear in their copy. That’s a crude test but so far it seems to work okay. If you cannot get a page to appear in a site: query for a small expression that only appears on that page (within your site), your page is not fully indexed.

At SMX Advanced 2007, Matt confirmed that you move pages from the Supplemental Index to the Main Web Index by getting more (value-passing) links to point to those Supplemental pages. At Domain Roundtable 2008, Matt said that if you have an existing network of sites and you add more sites to that network, you’ll need to get more (value-passing) links to avoid diluting your PageRank.

Call it a measure of PageRank strength, if you will, or PageRank share. That is, you cannot bootstrap your way to more PageRank the way you once could because Google is apparently requiring the PageRank to flow from site to site. That sounds like the domain-level PageRank idea that Matt McGee shared in response to a comment Matt Cutts made at SMX Advanced 2008.

There are some problems with the assumption of domain-level PageRank calculations. For example, Google still treats sub-domains as separate Web sites, and Matt Cutts seems to imply that Google thinks in terms of ’sites’. A site might be a host as defined in some old technical papers (a domain or sub-domain), but there are plenty of Web sites out there which are confined to special sub-directories (especially at many university and ISP Web sites).

Google could be calculating PageRank on a site-level but it doesn’t need to do so in order to throttle the growth of PageRank. We could certainly suggest other interpretations for Matt’s remarks.

For example, Google might increase the amount of total PageRank in its database (which would be a radical departure from the original definition of PageRank as a probability distribution in which all allocated PageRanks added up to 1).

Another possible interpretation might simply be that Google requires a site to receive inbound links from a minimum number of other sites (that already have some PageRank) before including that site in the pool of sites for which PageRank is calculated (in which case the average amount of PageRank per site or page would diminish naturally but at a slower pace than under the original model).

However you choose to look at what Google is doing with PageRank, the point remains the same: you need more (value-passing) links regardless of whether you add pages to your site or add domains to your network. Google will seek to strip domains of their PageRank if they change ownership (one way of “returning” PageRank to the otherwise growing pool of PageRank-eligible sites).

The question then arises of what happens to the anchor text? Can links pass anchor text if they cannot pass PageRank? That is, if you add pages to your site, will their internal links help you in any way? Just when we’ve finally got people to think about using descriptive anchor text other than “home” for their root URL, has Google pulled the rug out from under basic search engine optimization?

Of course, there are other search engines than Google, but Peanut Butter SEO is really confined to optimizing for Google.

We can draw several conclusions from this information:

  1. Google is suggesting that obtaining links from older, established domains will be more effective than buying older, established domains
  2. Google is suggesting that increasing your inbound links works better when you increase the number of (value-passing) linking sources
  3. Increasing the size of a site or network does not increase the site or network’s search visibility
  4. It is better to obtain more value-passing links than to attempt to direct the flow of PageRank through a site

In essence, every Web site will exhaust its potential for obtaining PageRank unless it obtains links from an ever-increasing circle of unrelated Web sites. In other words, simply playing to your niche strengths limits your site’s ability to increase its PageRank. You’ll have to acquire link recognition from sites that are not immediately relevant to your own site in order to increase your PageRank significantly.

Peanut butter SEO should help the community define new criteria for valuing and obtaining links. We may be about to enter an era where people talk about “PageRank pedigrees” and link lines when determining which sites should be used as linking resources.

{ 0 comments… add one now }