My experience is that Google increased the credit on “traffic” section of the website and makes it hard for SEOer to improve sites ranking.

One test I did on one of our company website:

1) Stop doing any back-links building for one week.

2) Only attract more real/natural traffic from different social media accounts.

3) After one week, I saw the ranking moved from #14 to #7 on the first page of Google.

Freelance would be a good sample of those sort of high traffic websites.

Google has ~1.5 million indexed URLs for the domain. Open Site Explorer shows a relatively small number of links & they aren’t from high quality sources. The site is full of thin content (each actual question/listing) and oddles of duplicate content (related listings) and as such a lot of boilerplate content.

I haven’t bothered to check, but I wonder if they are doing something clever with the related listings to try and avoid getting hit by the Google Panda algorithm. For instance, if they always list the same related listings on each URL – that’d be a problem. They could improve that by being ‘related’ by category, category+tags or category+tags+full text search. I’d then be tempted to sort the ‘related’ stories by a random number generator, to make sure that each page is going to list related by different items.

I’m sure the algorithm’s that Google use are very smart but maybe some relatively simple techniques like that slip through. After all, they aren’t duplicating an entire URL like you’d typically see with issues like www/non-www, HTTP/HTTPS, mixed URL casing, trailing slash issues and so forth.

Instinctively, I’d suggest that it isn’t a sustainable tactic for them. However with the rise and rise of StackOverflow – maybe it’ll last longer than we think. Granted StackOverflow is producing a lot of high quality, well curated content by experts but again – maybe the Google algorithms can’t detect it yet.

Thoughts?