In 2007 I solved an issue with one of my sites which could be classed as deoptimisation or overcoming the 950 filter. But I may have another issue now!
In June 2007 my ecommerce site started losing google traffic after the major update at that time. The site had been steadily rising in position and traffic for four years and had become one of the top six sites in its' niche (generating nearly £2m in sales) purely from lots of hard work optimising the site and a few good quality links. We'd never made big changes to the site in one hit, just gradual minor refinements/additions. I assumed we'd just tripped a filter somewhere.
Finding the cause was quite tricky though. I could see some product pages were no longer being found for terms that exactly matched phrases on the page and in the title (the product name), but we started getting traffic to those pages for other terms. So we lost traffic but gained some as well. The overall result however was a slight but noticeable loss. It looked to me like an over optimisation issue as google clearly didn't want the site to have traffic for terms it thought we were targetting, no matter how obscure.
In the months that followed I experimented with various de-optimisation approaches with no conclusive outcome. Meanwhile our traffic seemed to be constant but about 25% below our norm for the time of year. Then in mid-October 2007 things took a major turn for the worse and google traffic started disappearing at a rate that increased every day. Our main category pages were vanishing from searches by the day and eventually after a week our homepage went for our main term (although the site could still be found for the website name).
It was now clear I had a major problem to fix and not much to lose. I asked around and a few people mentioned some internal links I had in the footer to gift guide pages (21st birthday ideas and so on). I removed these links from the footer at the end of October 2007 and the next day when google had a new cache of my homepage, it was reinstated to its' previous position. As each day went by the pages returned and within one week our traffic had returned to full strength. But it didn't stop there, our traffic continued growing and eventually peaked at 20% more than we'd ever seen for a December (our peak business time - phew!).
The interesting thing is I was warned to remove the footer links two years before by an associate whose site had also been hit and recovered once the footer links were removed. So if the filter was in place two years before why hadn't we been affected before? It was like our site was okay until it hit a certain level of success, then came under closer scrutiny.
Anyway, at the start of 2008 I hired an SEO to see what other issues might be lurking. I found someone with a great reputation among people in the industry who I trust and who had produced excellent results in much more competitive niches than mine. He picked up on the following potential issues:-
1) Links from two other sites that I own to our ecommerce site
2) Duplicate content on the two other sites
3) Server location (was Canada, now US) - it's a UK site
4) Need for more quality links
He felt point 2 was not an issue for the ecommerce site (but may be for the other two sites), and point 3 may have some bearing on ranking but wasn't a major issue. Point 4 is always something to work on so that left point 1.
We run two affiliate sites (one focused on the same niche as the ecommerce site and one right across the board), but both offer price comparison between our products and those of our competitors (as an affiliate so we get commission if they buy from the competition). They deep link to product pages on our ecommerce site or that of our featured competitor.
I know linking between your own sites is theoretically a big no-no but the ecommerce site doesn't link back to the affiliate sites and the only difference between the competitor sites that they link to and our own ecommerce site is the Whois information. The sites are located on separate servers in different IP ranges. And these sites do generate visitors and sales for our ecommerce site so the links have a real and genuine purpose.
The SEO couldn't say for certain whether this interlinking would ever harm the ecommerce site or was harning it already. He didn't feel sure enough to recommend that I remove the links and suspected that it may well be helping while the affiliate sites had enough trust. He felt uneasy making a recommendation one way or the other.
Consequently we decided to leave the issues we weren't sure about and focus on improving the quality of the links to the site to build trust, move us up the rankings and hopefully make the site more resilient.
We chose some keywords and pages to focus on that had been around page 5 in the results to get them onto page 1, but within a couple of months we found the targetted pages were falling for the exact keywords we were targetting (while every other page remained unchanged). Eventually the targetted pages fell to page 10 or lower so we decided to call a halt to the link work. Google seemed to be resisting our efforts!
Over the course of 2008 this site has dropped slightly in google and the Alexa rank is also on a gradual slide. At it's peak in June 2007 it was ranked around 200,000 and now over 400,000.
At the moment there is nothing to clearly suggest google has a problem with the site as a whole and the only suspicious sign is one category page that has bounced in and out of the results all year like it's on the edge of a filter. Alot of the products on that page have similar names so there is alot of keyword repetition, but no more than some other pages that seem to be 'liked' by google.
Since November our traffic has been consistently 25% below what it was in 2007 but the economy and increased competition over the course of the year may be to blame for that, not necessarily a problem with google.
Sorry for the long post but I thought the history of how we solved our problem last year may help others, and maybe someone can help me figure whether we have a problem now and, if so, what we should do about it!