1 July 2014,
 5

Top Secret

Brian Dean’s list of Google’s 200 Ranking Factors is the most comprehensive list I’ve ever seen, but it’s missing one thing. One key ranking factor that is rarely talked about, but is just as important as any on-page or off-page SEO.

The Throttle Factor

But before I get into that, we first need to look at Google from the right frame of mind.

Mindset

At the end of the day, Google is a machine and all machines work in a certain way. Theoretically speaking then, there should be a formula to rank your site that should work across the board.

Why is it though that:

  • Some sites rank better than others with the same quality of content?
  • Some sites have trouble breaking into the top 100 whilst others have no trouble at all, given the same quality of links in the same niche?
  • Some people see an increase in rankings from a couple links whilst others don’t see anything at all from dozens of links?

The answers you’ll generally get for these questions are:

  • each link is different
  • each site’s on-page is set up differently
  • each keyword has a different level of competition

All of these are true, but is there more?

I’ve personally had sites receive the exact same links but react drastically different, so there must be something missing.

I thought about this and thought, “why is there no formula to rank”?

The answer is pretty obvious, Google or other search engines want to discourage SEO. If there was a formula for it, it would completely ruin their business and no one would ever pay for ads again.

Thus came the thought:

What if there is a 201st Factor that randomly affect how sites rank? This would surely throw all SEO’s off as it would make coming up with a “formula” that much harder.

I went out to test it.

The Experiment

Hypothesis

Google randomly throttles the rankings of some sites with no logical or apparent reason. Some sites are affected, whilst others are not.

Setup

To make this a completely fair and rigorous test, there needs to be some link manipulation involved (aka blackhat SEO).

We bought 12 brand new domains and set WordPress up on them. We decided to have 2 groups of sites:

  • Group 1: Home Repairs (7 domains)
  • Group 2: Home Construction (5 domains)

*there was no logical reason why we chose 5 and 7 domains.

Other details include:

  • Location: Canada
  • Each group consisted of a mix of .ca and .com’s.
  • Each site was setup exactly the same way (same title, H1 tag, plugins etc.).
  • All of them used the Genesis Theme

Execution

We sent identical anchor texts from the exact 5 same PR sites to each Money Site (see diagram below).

Secret ranking experiment diagram

This meant that each money site had the exact same quality of links and anchor text coming in.

Results

Results were very interesting, as we saw a very large variation in rankings.

Home Repair Niche*

*This test was done prior to the Panda 4.0 update.

Prior to any link building, we already saw a huge variation in results (some KW sitting on page 4, whilst others on page 11).

home_renovation_before

After some link building, we saw an increase across most keywords, however the links truly highlighted the winners and losers.

home_renovation

The winner of the group of 7 domains was ranking 9th for 1 keyword, whilst the worst performing site was ranking 106.

Home Construction Niche*

*Right after we ran our first test, panda rolled out so we decided to do a couple more. This test was done in June 2014

Here are the results:

Home Construction Niche 1

As you can see, there is clearly 1 winner which Google clearly likes a lot more, but for no apparent reason given that the sites were all set up the same were the links.

One thing to note with this test is that post panda, a lot more sites are not ranking at all.

Conclusion

Our hypothesis was correct in that there is a throttling that causes some sites to rank and others to not.

In addition, we can note that:

  • There is no logical or apparent reason that we can currently see that Google uses to decide which sites rank or not
  • In both these tests, the ranking domain happens to be a .ca TLD (we’re targeting a local term in Canada). Perhaps this is a factor, but we have seen other .ca’s not rank well. .CA domains may help domains rank in Canada but it looks like this random factor is applied to all TLD’s.

We’re still finalizing some other tests before we put them up, but each test is indicating that there is some additional factor that is built into each site that causes it to rank (or not rank).

For the latest on throttling and the latest SEO news, check out our test lab.

5 responses on “Google’s Secret Ranking Factor that You’ve Never Heard Of

  1. Troy says:

    Very cool and very interesting concept “throttling factor” – please update your findings as they come in – awesome that you setup an extensive testing trial for this!

  2. Rusty says:

    Hi,

    Not a authority on SEO, but could the viewer be the random factor, as in perhaps more time spent on this site. I guess you can check this…

    Also, it raises a question of mine regarding having a 3rd party (fiverr style) provide pageviews from random IP’s would help rankings on sites.

    CHeers,
    Rusty

Leave a Reply

Your email address will not be published. Required fields are marked *