Google has just announced that their “Panda” rankings changes, first launched in the United States in late February and rolled out to English language indices internationally in April, have now launched internationally in all languages other than Chinese, Japanese, and Korean.
Their post notes that for non-English indices, this change impacts 6-9% of queries (vs. the 12% the original US English launch). This launch also includes a few minor changes to the English version of Panda, but shouldn’t have a substantial impact.
In their post, they say:
For many months, we’ve been focused on trying to return high-quality sites to users. Earlier this year, we rolled out our “Panda” change for searches in English around the world. Today we’re continuing that effort by rolling out our algorithmic search improvements in different languages. Our scientific evaluation data show that this change improves our search quality across the board, and the response to Panda from users has been very positive.
I talked to Google about the change and they reiterated that searchers overwhelmingly have found the Panda-based changes to improve their search results and that those impacted should evaluate their sites objectively for quality and unique value.
This year, Google has focused on identifying sites with a large number of low quality pages as part of their overall goal of providing the best possible search experience. The Panda updates have been evolutions of algorithms that increasingly detect this and lower those sites in search rankings. I often describe it as somewhat like the Adwords quality score, which uses a number of factors to assign an overall score to a site. Since we’re talking about algorithms with many inputs, there’s no one thing that can cause a site to lose rankings due to Panda. Rather it’s an accumulation of factors.
Imagine you’re eating an ice cream sundae. The ice cream is delicious and creamy. It’s covered in the best hot fudge sauce you’ve ever eaten. The whipped cream is freshly made from scratch. On top is an OK-tasting cherry. How do you feel about the sundae overall? Pretty good? Excellent even?
Now, imagine another ice cream sundae. It’s made with that blech tasting cardboard-and-ice style ice cream. The hot fudge is missing entirely. And the whipped cream is that scary fake stuff from a can. The cherry, however, is quite good. How do you feel about this sundae?
The maker of the second sundae might ask why it is you’ve put down your spoon and are edging back to the sundae #1. But the cherry is better! He might say. A Rainier cherry from Washington state! How can you like the other sundae better when it has such a sub-par cherry? Mine has all of the same parts as that sundae! Well, except the hot fudge!
Google is taste-testing ice cream sundaes and offering searchers the ones that are the best holistically. I mean, they’re not literally eating ice cream. (Actually, they are, but not as part of this metaphor.) I’m just saying you can’t look for one specific thing to fix and you can’t compare one specific thing on your site to that same thing on another site. There are too many moving parts.
The History of Panda
This post covers all of the Panda updates leading up to this one. That post also recaps the articles we’ve published here with advice and impact. The last update was about three weeks ago and was fairly minor.
Advice If Your Site Has Lost Ranking
Panda seems to be focused primarily on unique value and user experience. I gave a long interview to Eric Enge about this not too long ago. If your site has lost search traffic due to Panda, take an objective look (or better yet, have someone else take an objective look) and ask:
- How does the content quality compare to other pages on the web about the same topic? Is the page the most valuable and useful content about the topic?
- Do multiple pages on the site answer the same problem/focus on the same basic task? It’s one thing to have separate pages on “best chocolate cake recipe” and “best pumpkin pie recipe” and quite another to have separate pages on “best chocolate cake recipe” and “ideal chocolate cake recipe”.
- Is the content primarily syndicated or aggregated from other sources? If most of the content isn’t original, Google’s algorithms might give the site a lower “quality score” (in quotes because I totally made that up that way of looking at Panda — I’m not saying Google internally is using the concept of a quality score) to better ensure that the original version ranks.
- If the content is unique, does it completely cover the topic in a credible, useful way or is it shallow and barely scratch the surface?
- Does the user interface design and navigation make engagement easy or are things cluttered and make it difficult for visitors to find what they’re looking for?
- Are the site design and goals user-focused or revenue-focused? It’s absolutely fine (and generally necessary if you’re running a business!) to ensure that your site makes money. But if the goals you keep in mind when designing the pages don’t take into account how well the visitor can get what they need (an answer to their question, ability to accomplish their task easily) and only are based on getting what you want from users (ad views or clicks, for instance), the user experience of your pages might not be ideal.
Since Panda is based on an overall ice cream sundae score, you likely won’t see rankings improvements right away once you make changes. Google periodically recalculates these scores (just like they periodically launch a new version of Panda with improved signals), so after you make changes, you’ ll have to wait for Google to recrawl the site so they can take note of the changes (you can check the cache dates of your indexed pages to see if Google has recrawled them) and then you’ ll have to wait for one of Google’s periodic scoring recalculations (which so far seem to coincide with Panda algorithm updates).
Some site owners who have made substantial changes based on the bullets above have seen positive results (I’ve worked with some of them and seen the analytics myself), but a recovery isn’t likely overnight.