Google's brand/trust algorithm

Easy to figure out:

Branded searches over time. Most authority sites have large amounts of branded traffic. Some for example have more branded searches than non-branded. Some are also in the 100,000's and more.

This is caused in part by offline branding, but also online brand recognition over time. To manipulate this, consider having people search for your brand in all offline collateral and encourage it through online collateral.

To attempt to cheat this system, use automated query tools with random everything and mask IP's, locations and what not to confuse their systems.

The benefit is boosted non branded traffic. This is because it assigns some level of trust based on users and how they search. Google is not that smart more 'sneaky'. You just need to think how they would think. All their systems can be exploited.

Examples of branded queries are queries with www.domain.com, domain.com domain com, etc.

I used to use fake SERP traffic tools and it USED TO work great, but lately its had no real effect I can see. I think Google is looking for better quality traffic and even the exchanges get their IPs abused and it lowers the quality too much.

The other part to this equation is the 'punter can't find the library book' as the Google librarian has moved it again. This is where the user would query: brand + non branded query or term and even add location data e.g. kijiji toyota toronto.

Enough of this kind of data and Google can figure out what users want. As for your IP pools, that would be a problem yes. I think someone tested this a while back with Mechanical Turk, and it worked. They might have some fallback or redundancy built in, to protect their systems, but it's still an exploitable ranking method. I can easily list a handful of signals they look at that most overlook