An SEO Experiment Gone Awry (Mea Culpa)
Today I am writing the most difficult blog post I’ve ever authored. Why so difficult? Because something I did has made some people very unhappy. So here I will state my case and lay myself before the mercy of the jury. I do not know if it will help, but at least the information will be available for any and all to read.
For those of you who know me personally, you can attest to the fact that everything I do is on the “up and up”. I’ve been called honest to a fault on numerous occasions. I value my reputation like I value those I love and cherish. In the spirit of honesty and transparency, I present the following.
Google Panda and an SEO Experiment
Several months ago, I first learned that Google was working on a algorithm update that was meant to penalize low quality and duplicate content online (now known as “Panda“). Needless to say, I was intrigued. As an SEO and online marketing practitioner professionally, I had witnessed sites like eHow manage to leap to the top of the rankings for nearly any search term you might enter into Google. On the other hand, I had syndicated blog posts from friends and colleagues here on Return On Now… content that would now be considered “duplicate” by Google. I was worried my SEO would take a negative impact, so I kicked off an experiment to test just how hard the update would affect me.
Structuring the Experiment
To perform an SEO experiment, you need content. In this case, I needed duplicate content. But I wanted to see just how aggressively Panda would penalize it. So I decided to design two test sites, one with essentially all duplicate content, and another with a handful of original pages plus some duplicate content as well. Since this was to be a short term trial, I built two WordPress sites using freely available standard templates and found a plugin that let’s you create posts from RSS feeds. Bingo! My experiment was ready to go.
Then I started thinking about what topic areas might be rich with RSS feeds to run the experiment.
I had already purchased “supplymyhobby.com” for an etail business idea I kicked around in 2010, and it was sitting there unused, so I chose it for the first site. Obviously, hobby content is widely available on the web, so I decided to make it the “all duplicate” site.
I also selected a new domain focused on back pain, an ailment I’ve suffered from since a car accident in 1994. I figured, since I am in the demographic of who would want to read this sort of content, I could hack together several pages of original content to supplement the duplicate material. In this way, I could test whether Google Panda would “slap” a site harder based on how much of the content is duplicate.
On both sites, I wanted to be sure Google indexed them as real sites and not experiments, so I did add some more promotional content. One asked for what hobbies readers want to see on the site. The other had a few pages of product reviews for back relief remedies.
Populating the sites
I built both sites first, and then configured the autoposting plugin. At this point, I actually had second thoughts and nearly scrapped the whole idea, but I figured that this was a short-term trial. I could set it up, get the sites live, ensure they are indexed in Google, and then watch the traffic trends until a few weeks after the Panda rollout. Then I could draw conclusions, turn it off, and integrate the learnings into my ongoing SEO work. The key phrase here is “turn it off”, which is where it gets hairy.
Now, I understand the slippery slope of using content from other blogs, There have been many debates about what a copyright means online vs. in print, what rights authors have, and what attributions are required for copied/shared content. As you’ve seen here on RON many times, I frown heavily upon stealing content for personal gain or other financial reasons. I would never, ever steal someone else’s high quality content for the sole purpose of taking credit for it or making money in a shady fashion. This is an important point, and one I will come back to shortly.
To get the sites indexed in Google, I had to connect it to some RSS feeds. Since I needed to get the site live as fast as possible to build a little momentum pre-Panda, I hurried to connect some RSS feeds that were serving up quality on-topic content and turned the sites on. I also set the content autoposting plugin to append two things on each post, the official name of the source blog and a link to the original content. I figured, at least I’m establishing backlinks which wouldn’t have hurt the original sites in any way pre-Panda, and I’m being upfront about what sites actually created the content. I did this because I am no content thief, and I would never do such a thing otherwise.
In this process, I did overlook one important thing on the WP template – I didn’t change it at all. That was an outright mistake, because they came with verbiage claiming that all material on the site was copyrighted by the site itself. Shame on me for the oversight, because I knew full well that some or all of the content would be written by others on their own sites. It has been suggested that I should have simply asked for permission, and I can’t argue that point. In my haste to make the experiment happen, thinking no one would even notice, I did not do so. Again, shame on me.
Once the sites were live, I did not promote them aggressively. Upon launch, I did link to them from another website or two, ping Google directly, and bookmark a handful of pages and posts. Basically, just enough to get them indexed. My goal was not to build traffic, grow the sites, or create some sort of business. It was purely academic, and once indexed, only organic search traffic needed be measured to draw any conclusions.
Let’s keep this part short. What did I learn about Google Panda in this experiment?
- As we saw across a variety of sites, the penalty for duplicate content was swift and severe. Traffic essentially fell off a cliff back in February when it went live.
- There was a marked difference in negative impact between the two sites. The faucet was nearly turned off completely on the hobby site, while it merely took a downturn of >50% on the other site.
The second observation was especially enlightening, because it showed that Google weights its rankings according to amount of duplicate content, not an across-the-board slap for having any of it. This is exactly what I was hoping to see, because now I don’t have to worry about my organic traffic completely drying up on Return On Now.
Great, experiment over. Back to business….Not so fast buddy.
Finish What You Start
I made a crucial mistake at that time. I neglected to turn off the sites as planned.
I could spew a littany of excuses including adding a new client for my SEO business, rolling out a major new website for a local tech company, and having to deal with some personal stuff that distracted me elsewhere.
But bottom line: the sites lived on.
Earlier this week, a Google Alert came to my In-Box that really caught my attention. One of the bloggers whose content I used in the experiment (In Stitches) had seen SupplyMyHobby and was rightfully upset. I knew immediately that leaving up the site had backfired, and that I now had created a mess for myself. The old adage “finish what you started” came to mind, and I realized I had completely dropped the ball.
You see, In Stitches is a GREAT blog. They have a good quorum of regular participants on the blog, and the author (Pam MacKenzie) has built a wonderful online presence within the knitting community. I respect her work with the highest regard, which is part of the reason I used the content in the first place. I feel the same way about every one of the blogs that populated SMH during this experiment.
This Google Alert pushed me to immediate action. After reading her scathing blog posts directly, seeing words such as “plagiarize” and “stealing”, I immediately opened my hosting FTP account and deleted the two sites completely. Of course, this was over 2 months too late, but I removed them from the web without delay and also deleted the Google Analytics accounts I used to track the experiment. I wanted no semblance of these websites to remain live online for any reason.
As I mentioned earlier, I am 100% dedicated to honesty, transparency, and taking responsibility for my own actions. The next step I took was to send an email directly to Pam with a full apology and an attempt to explain what I was doing. But you can only say so much in an email, which is why I am posting this live on my REAL blog for the whole world to see. Nothing to hide here.
Picking Up the Pieces
Ms. MacKenzie continues to think I am a scammer.
While I can understand why she might think so, I implore you to take this into account – if I were actually stealing content and taking credit for it, why in the world would I have attached it to a hosting account that lists my real name? Just search for me online and you’ll get a whole page of links to my various blogs and social media profiles. I bare myself to the world in full color, without editing or filtering. I am who you see online, in person, and in writing.
It’s known that scammers make up fake IDs, names, and contact information so that you can never actually trace their footprints. Do the math for yourself, and you’ll see that this is a severe misunderstanding that blew up in my face in the worst way possible.
I screwed up, and I apologize from the bottom of my heart to every one of the bloggers whose work I used in my experiment without asking for advanced permission. It was wrong, and I’m done fielding isolated SEO experiments altogether. It’s far too risky to touch again.
Lesson learned. I hope you can understand. Namaste.