Haven't posted in several months because I got disgusted and stopped monitoring our sites. In brief recap, we have three sites, no thin content, oldest site from the mid 90's, newest site around three years old. The two larger business sites got hit on February 24th and again around April 10th, losing 70% to 80% of Google traffic. The smallest site actually benefitted on February 24th, maybe a 30% gain, then got killed around April 10th.
The smallest site, which drew around 500 visitors a day from Google pre-Panda and is back over 400 now after averaging less than 200 a day since April 10th, consists of around 20 plain HTML pages. As my hobby site, it has no commercial application, no ads, no links to other sites that I can remember, other than YouTube. After it got hit by Panda, I deleted a small blog that I hadn't updated for a couple years, but that was back in April.
Other than that and a single new page I added to the site last month, I did nothing. No HTML tweaks, didn't rework the navigation, didn't change the aesthetics, run a new spelling or grammar check, didn't make the site more "trustworthy." Didn't think about the user experience, run focus groups, add social content, none of the stuff Google uses to explain what they call quality.
The Google traffic is still down 20% year-on-year, but given that it had been down over 120%, I call that a rounding error. It is sort of interesting that the keywords for which Google is sending traffic appear to have shifted radically, but I'd want around a month's worth of data before trying to make a comparison like that.
The improvement was not a step function like the original penalty, rather it's been ramping up for the last 10 days. I don't know if that's just the new take spreading to different Google data centers, if they've been slowly backing out a screw, or if it's the off-site changes. It's not a page-by-page thing, they are all rising more or less in unison.
So here's the weird backlink thing. While I didn't make any changes to the site, I did make some changes to our much larger business sites that linked to it. Since it's my hobby and I like driving traffic to the site, I used to have lots of links to it from the two sites that were drawing over 10K visitors a day pre-Panda. Those links never got many clicks, but they contributed a lot to the hobby site PageRank.
During a navigation redesign of one of the sites a couple weeks ago, I dropped those links from all but one of the pages. When I saw some improvement on the hobby site last week, I came here and read that Panda may be updating more frequently (or not). So I removed most of the links to the hobby site from our other business site as well.
So I'm left wondering if the hobby site, which was taken down by Panda in the cycle after the business sites, is now recovering because it's now less associated with them. All three sites are on the same server, same site owner registered, if Google looks at such things, it would be clear they are all owned by us.
The hobby site has always had the worst bounce rate, time on site and pages per visitor of any of the sites. It has the lowest PageRanks, fewest incoming links and the lowest quality incoming links (mainly from hobbyist forum posts).
In a side note, the two business sites show no inkling of recovery, in fact, they are slipping further at a time of year they would usually be trending up. Unlike the pages on the hobby site, many of the pages on the business site have been through a full editorial process (editing, proofreading), the business sites have been around much longer, have many, many times more incoming links, higher quality links, etc.
