I apologize to Google. They still like my blog.
This blog’s numbers plummeted as per Webmaster Tools, here and here you find everything you never wanted to know about it. I finally figured that my blog was a victim of Google’s latest update Panda 4.1. Sites about ‘anything’ had suffered, and the Panda rollout matched the date of the onset of the decline.
Other things happened in autumn, too: I had displayed links to latest WordPress blog posts on my other websites, but my feed parser suddenly refused to work. The root cause was the gradual migration of all WP.com blogs and feeds to https:// only. Only elkement’s blog had been migrated at that time; our German blog’s feed was affected two months later.
Recently also the German blog started its descent in impressions and clicks, again two months after elkement’s blog. I pondered about https URLs again – the correlation was too compelling. Then suddenly the answer came to me:
You need to add the https URL as an additional site in Webmaster Tools.
It was that simple. All the traffic I missed was here all the time – tucked away in the statistics for https://elkement.wordpress.com. This also answers the question I posed in my last Google rant post: Why do I see more Search Engine referrers in WordPress stats than clicks in Webmaster Tools? I had just looked in the wrong place.
I had briefly considered the https thing last year but ruled it out as I misinterpreted Webmaster Tools – falsely believing that one entry for a site would cover both the http and the https version. These are the results for both URLs – treated like separate entities by Webmaster Tools:
Results for http : // elkement.wordpress.com – abysmal:
(Edit: I cannot use a link here and have to add those weird blanks – otherwise WP will always convert both URL and text to https automatically even if the prefix is displayed as http in the editor.)
Results for https://elkement.wordpress.com – better by a factor of 100: Popular pages were the first to ‘move’ over to the https entry. This explains why my top page was missing first from http pages impressions – the book review which I assumed to have been penalized by Panda as an alleged cross-link scam. In full paranoia mode I was also concerned of my adding random Wikimedia images to my poetry.
You have read a post in my new category Make a Fool of Myself. (I tried to top the self-sabotaging effect of writing about my business website being hacked – as a so-called security expert.)
Yet the theory was all too compelling. I found numerous examples of small sites penalized by Panda in a weird way. See this discussion: A shop’s webmaster makes a product database with succinct descriptions available online and is penalized for ‘key word spamming’ – as his key words are part of each product name. Advice by SEO experts: Circumscribe your product names.
Legend has it that Panda was named after a Google engineer. I figured it was because the Panda is so choosy, insisting on bamboo
eucalyptus (*), just as Google scrutinizes our sites more and more. (*) One more theory I got wrong, now edited! Thanks to commentator Cleo for pointing out the mistake.