This year has taught me some very important SEO lessons. I’m going to share my most embarrassing one with you now.
Crouching Panda, Hidden Dragon?
In February of this year (2011), Google rolled out their “Farmer” or “Panda” release, which aimed to penalise cruddy content on the web (the fluff that’s designed only to generate low-quality links).
On March 1st, we redesigned Web Design from Scratch, using a lovely new WordPress theme, designed by my creative director Jimmy Hughes and implemented by production director Dan Johnson. And we were very pleased with ourselves!
Until this happened!
It was pretty clear what was going on. Google had somehow mistakenly evaluated our site as having low-value content, and had penalised us in the search results.
We set to work trying to figure out what we’d done wrong – and what to do about it.
- We drafted in help from one of the world’s top SEO experts.
- We went through and fixed dozens of broken links with 301 redirects.
- We scratched our heads – a lot.
Nothing made any difference.
Until one day recently, I was doing some research on clickthrough rates from search engines, to help my course students understand better the link between SERP position and clicks.
I noticed from Google Webmaster Tools that the clickthrough rates for some of our most popular pages had gone down. But our rankings had not. It was the clickthrough that had caused the drop in our traffic, not our rankings.
In other words, it wasn’t Google slapping us. We had found a way to slap ourselves.
But What Caused the Self-Slap?
I’m going to show you a sequence of pictures.
Can you spot the difference between these two?
Here’s the first one…
And here’s the second…
Did you spot the difference? It’s very subtle.
The second snippet featured the original date the article was written – back in 2006!
That’s the difference, and that’s what caused our search engine visits to halve.
When someone is looking for insight into modern web design styles, they don’t want a page that’s five years old. It will seem out of date. That’s just a fact of life.
We hadn’t had dates on our posts before. Some of our busiest articles – the ones with the most inbound links that get great rankings – were written a few years ago. But previously they didn’t publish a date, people read them, loved them, and kept linking to them, so their rankings only got more consolidated.
Overnight, we add dates, and traffic nosedives, by why did it take so long to spot?
It wasn’t until we looked at the data at different way. And I needed to realise that number of searches and search rankings don’t automatically turn into visits. There’s another crucial step, which we often overlook, and that’s the appeal of the SERP snippet, as I explained in yesterday’s videos.
Of course, we couldn’t be sure that displaying the dates was the significant factor without testing. So Dan hacked WordPress, which now shows the date – but only for 3 months – after which the date is automatically removed (unless we want to keep it). See Dan’s tutorial on how he removed the dates.
We did that last week. Here are the results so far.
This is just our organic search engine traffic. You can see that we’ve recovered from about 1500 daily visits from search to 2190 yesterday.
We need it to increase by another 50% to get back to February’s levels of traffic (about 3500 daily visits), but we’re on our way.
One of the things the Panda release was supposed to do was also to penalise pages that did not get good clickthrough rates from search results. So it may be that we did actually pick up a penalty from Google after our CTRs suffered initially, which should also recover in time.
Here’s the list of our most-clicked pages from Google Webmaster Tools (which is always a few days behind, so this doesn’t include the last 3 days.
The dip in the graph represents the weekend. But look at the far right column: Change in Avg. position. The green numbers represent pages that are ranking better than they were one week earlier. Red signifies a worse ranking.
Our top 2 queries have gained 4 and 5 places respectively, and the overwhelming change is very positive.
This supports the theory that Google rewards pages that get good clickthrough rates.
And here’s an update a few days later. The results continue to improve over the week since the change.
Update, December 2011
The good news is that organic search traffic has recovered gradually, so almost to the same level as this week last year (16,000 visits against 17,000 last year).
The growth since June 2011 has been startlingly consistent. It makes me wonder whether Google is deliberately turning our ranking back up.