Lab Report: Post Panda

We had the interesting experience of being smacked down by a Google algorithm update last month. We’ve never had this happen before on any of our sites, including any of our client sites. Algorithm changes are usually good for us, for the simple reason that we always focus on good content and visitor usability, eschewing all questionable SEO tricks.

So when our lab site, FreshPlans, lost half its traffic from Google (not bing or Yahoo) between October 13 and October 14, we saw it as a good opportunity to figure out what had gone wrong and what to do about it.

Nah. We tried hard to figure out what we had done wrong. I often find that our clients can’t see anything wrong with their websites when we can see the problem immediately (a good reason to hire people like us to do a site analysis for you), so we may just be overlooking it. But we did scour the site and listen closely to all the rumors about the algorithm change:

• We were out of the country when the change took place, so our daily postings had gone down to no new postings for a week. It seems excessively strict to chop someone down for going three days without new content, but we got right back into the daily schedule as soon as we returned.
• We have lots of outbound links at FreshPlans, since we direct people to good teacher resources elsewhere. We’re definitely curating, not link farming, but we went ahead and did more posts without outbound links.
• We have some affiliate ads on our lab site, and we do play around with them occasionally in an experimental way (lab site, remember?). We don’t feel that our ads are too aggressive, though. We saw no way to scale them down, short of removing them altogether. We had just installed Adwords, a Google product. but it’s hard to believe that Google would smack us down for using their product. We backed off on including affiliate links within our text — no lists of books, even though that’s something teachers appreciate.
• We heard a rumor that the change had penalized companies that didn’t have geographic information, which FreshPlans did not. We added a street address to the Contact page.
• We also heard that bounce rate was involved, and FreshPlans had a high average bounce rate. It’s still higher than we’d like, though it had gone down somewhat until the adventure of the vampire babies. I’m going to write more about bounce rate tomorrow; it’s too big a question to go into here. Suffice it to say that we worked on decreasing our bounce rate.
• Another rumor said that it was not so much websites as keywords that had been affected — that is, sites that focused on keywords with a high level of black hat activity had gotten smacked. That makes sense to me, since I know that terms like “free lesson plans” and “printables” are rife with bad SEO behavior. We don’t focus on those terms, but they do come up naturally in the stuff we write about, and “lesson plans” was our primary keyword. Again, we backed off. We usually send people elsewhere for basic information and give specific ideas on how best to use that information in the classroom, but instead we’ve been providing information (biographical data, for example) even if it’s readily available elsewhere online. Obviously, it’s thoroughly written and researched original content.

Obviously, these aren’t discrete, scientifically tested changes with all the variables controlled. With the high level of seasonal change we see at FreshPlans, and knowing that it usually takes a couple of weeks for changes to show fully, it wasn’t practical to test each possibility. You can test user response efficiently, but not robot response. So we made changes one at a time, and didn’t see any immediate striking effects, and that’s about all we can say.

So did all this make a difference for us? Here’s the traffic for the relevant time period, compared with the previous year:

Here are the raw numbers for weekly Google organic traffic during the relevant time period in 2011, compared with the same time frame in 2010:

As you can see, 2011 showed a severe drop following the algorithm change between weeks 1 and 2, where 2010 showed a rise. The following week, Halloween, both years saw a drop as teachers quit looking for seasonal lesson plans and began using what they’d chosen. The chart for 2010 then resumes its steady upward trend, while 2012 continues to bob up and down.

Naturally, we don’t like this. But what can we learn from it about responses to Google algorithms?

• While we did lose significant traffic, we’re still up significantly from last year, and Google continues to bring us about two thirds of our traffic. What’s more, we may not be seeing the steady week over week increases we’re accustomed to, but we’re also not seeing a complete loss of traffic or a complete lack of growth. We conclude that a site can recover from a smackdown.
• Google uses hundreds of different factors in their algorithm and makes hundreds of changes each year. We’ve always said that the math doesn’t favor the idea that you can focus on one little trick and expect any results. We see no indication that any of the narrowly focused changes we made had any effect. Even though we continue to believe that FreshPlans is a high quality website, we still have to take the same approach we always do when people come to us for help with a Google algorithm smackdown: add good content, improve usability, and watch for gradual improvement.
• Right before the algorithm change, Google search had jumped up to over 70% of our traffic. We make sure, in our business, not to let any one customer provide more than 30% of our income, since we know that nothing in this life is certain and we don’t want to have a crisis if we lose one customer. We didn’t draw the same conclusion about website traffic sources, but now we think we should. If you see nearly all your traffic coming from one source, get to work on those other sources.

Probably the most important outcome for me in this experience has been a greater degree of compassion for people whose sites have gotten hit. I’ve been sympathetic, of course, especially for people whose income has been affected, but there has definitely been an underlayer of “Come on — treat this as a wake up call and get that site cleaned up” in my mind when I see people getting agitated about algorithm changes. We’re proud of FreshPlans, and it felt like a game of chutes and ladders when we were affected by the update. If you think we’re being blind to the flaws of our website and have advice for us, I’d love to hear it.

In any case, I now know how it feels. We’re working right now with a company that got smacked by the original Panda update, and I can see why. We believe that we can help him. We believe that FreshPlans will bounce back, too. But we can see the process through the eyes of the unfortunate site owners now, and that’s probably a good thing.