Google's "Brand" Update - Part 1

This week many webmasters have noticed the effects of a relatively large Google algorithm update. In fact, many of our clients that have a presence in the US will already be familiar with this. Rolled out in the US in January, "Vince" (so called because "Vince" is the name of the Google Engineer mainly responsible for the update) was jumped upon by the SEO community as a Google algorithm update that explicitly favoured large household brands. Vince has now been live in the States for almost 6 months; Google has had plenty of time to iron out any problems caused by the update there, so it's safe to assume that now they've chosen to roll it out to the UK as well the change is here to stay.

It took some quiet analysis and a comment from Matt Cutts to establish that the update was not, in fact, centred explicitly around brands, but was a major adjustment to the algorithm that placed an increased emphasis on the concept of "trust".

Here in the UK we seem to be going through the same furore again. In the first of a series of blog posts attempting to demystify exactly what "trust" is, and how the Brand Update works, we first discuss "TrustRank".

What is TrustRank?

Originally described in a joint paper between Yahoo and Stanford University (PDF download), TrustRank is a link analysis algorithm that takes as its central premise the idea that a small set of hand selected websites can be said to be the source of all "trust" on the Internet. It’s similar in nature to the more widely known PageRank. Like PageRank, TrustRank incrementally falls the further one moves away from the source of the trust. But whereas PageRank originates and flows from all pages on the web that have been discovered by a given search engine, TrustRank originates and flows from a small, select group of "seed" pages that are hand chosen as being trustworthy sources of link citations. In this model a site that is not linked directly or indirectly from any trusted site receives no TrustRank.

You may have read about Google having "well over a hundred" factors that make up its algorithm; TrustRank, or something similar, is one of these.

In the Vince update, Google seems to have placed much more emphasis on TrustRank as a component of its algorithm in its continuing quest to eliminate web spam from its search results.

Why would TrustRank only affect big brands?

There are two answers to this. In part, it doesn’t; closer to the truth is that it affects a number of sites including big brands, but that it’s big brands that most commentators notice and/or fixate on. Undoubtedly however the sites that have gained most from the algorithm update tend to be household names. This makes sense when you consider how TrustRank works; well known brands, by virtue of their very size and reputation, are linked to by trusted websites more frequently than smaller outfits.

Take the BBC for example. It maintains a stable of stock websites for each topic area that are included with relevant stories in the right hand panel under the heading "related internet links" (no other external links are included in BBC news stories). These related internet links are carefully vetted by BBC editors and typically comprise big brands, government agencies and the like. In addition, when a news story features a particular company, it gets a link from the story. And which companies are most often written about in the news? That’s right; big brands.

This doesn’t apply to just the BBC, of course. All sites, regardless of their TrustRank, tend to link to big brands or other well known sites more frequently than small sites. By the law of averages, it follows that big brands will also receive more of a share of the available TrustRank on the web. PR begets links, which begets PR. Thus a well known brand will tend to have its trust continually reinforced by more and more links from trusted sources.

What are the sources of TrustRank?

We don’t know if the BBC is actually an original source of TrustRank, although as bets go this is a fairly safe one. Even if it’s not, it’s likely to have a lot of TrustRank because it is linked to so widely from other sites. Unfortunately there is no published, publicly accessible list of seed sites for any search engines implementation of TrustRank. While there are some obvious candidates (e.g. the BBC) and some possible ones (I would speculate that Wikipedia might be one despite the presence of nofollow on all external links), inevitably there will be many we’ll never know about.

What's clear is that Google's search results have suffered at the hands of cheap, low quality links for years. Numerous attempts to devalue this approach to SEO have been and gone without notable success. Now Google seems to have taken a different tack; by improving the importance of high quality links, the relative importance of low quality links has been diminished.

---

Brand Update Posts

• Part 1 - What Is Trustrank?
• Part 2 - Could Search Behaviour Determine Rankings?
• Part 3 - Domain Authority

 

Comments

[Phil]

Interesting. Are their any tools or queries that could used to highlight TrustRank or BrandRank for a url e.g.

site:bbc.co.uk "RELATED INTERNET LINKS"
site:wikipedia.org "References"
site:wikipedia.org "External links"

Note: Google has blocked the allintext:"some-text-here" command.

Cheers

Phil.

[Adam]

Hi Phil, thanks for dropping round.

No there is nothing that will display a score for TrustRank or "BrandRank", not explicitly.

Those site: queries will simply list every page on the site in question, say Wikipedia, that contains the specified string, e.g "External Links" - obviously in Wikipedia's case that would be almost all of them.

If you want to find out how many links a particular site has from Wikipedia, you'd have to combine a link: or linkdomain: operator with the site: query. But search engines aren't a particularly accurate source of this type of information, you really need to pay for detailed stats from somewhere else, or even better just crawl Wikipedia yourself.