Google Panda Update: Could Inaccurate Google Data Be Costing Sites Traffic?
Late last week, it was discovered that Google had rolled out another version of the Panda update earlier in the week. Industry voices have dubbed the update “2.5”. Google dubbed it “one of the roughly 500 changes we make to our ranking algorithms each year.”
Did you notice a drop or increase in traffic in the past week or so? Let us know.
SearchMetrics put out lists of the top winners and losers from the update. Some sites were surprising, some weren’t. Interestingly enough, eHow and EzineArticles, which were previously “pandalized” were not on the loser list this time. EzineArticles would not offer comment, and eHow (Demand Media) told us that they’ve been pleased with the results of a massive content clean-up initiative they’ve implemented this year.
Another previous victim, HubPages, was even able to make the winners list this time around. Some of the more surprising “losers” were press release distribution services Business Wire (which actually just patented its SEO strategy) and PR Newswire, and tech blog TheNextWeb. There have been some questions raised over the accuracy of the SearchMetrics data, however.
“I’m glad to say we had a good summer as far as traffic is concerned,” Rod Nicolson, VP User Experience Design & Workflow for PR Newswire tells us. “We’ll continue to monitor closely, but so far we’re not seeing any unusual changes to our traffic due to Panda 2.5.”
TheNextWeb Editor in Chief Zee Kane tells us, “We haven’t noticed any effect right now but we’re still digging in. Will hopefully know more over the course of the next week.”
We’ve reached out to SearchMetrics for comment, but are still awaiting a response. We’ll update when we receive one.
Update: Here’s what SearchMetrics tells us about its data:
We monitor a selected and representative set of keywords for Google (in several countries) once a week and analyze the search results pages for these keywords. One of the main indices we calculate from this is the Organic Search Engine Visibility. This is a culmination of figures collated from search volume (ie how often people are searching for a keyword or phrase) and how often and on which position (ie what position on a Google results page) a domain/web site appears. Add them all up (plus some more math applied) and you get the performance index – an estimate for how visible a site is on Google in a specific country.
The basis for our analysis is a local keyword set for every country we analyze. Our values are local, that’s why we can give you an overview over the SEO and SEM visibility per country. The keyword sets are representative and varied between some hundred thousand and 10 million. The keyword sets are extended every month with new keywords added and irrelevant / outdated keywords deleted.
While we track millions of keywords, we obviously don’t track every single keyword that is searched. We can be viewed as providing a very good indication of underlying trends. However, results can be off when, for example, a web site has only a very small visibility and is ranking for a small number of keywords or a higher percentage of the keywords a domain is ranking for is not included in our keyword set.
Please note: Visibility is not the same as traffic. Further, sites that are listed among a ‘losers’ list may still generate traffic from other sources and can still potentially continue to prosper. Our data can only be used as a trend for search engine visibility on Google. But Google isn’t the only traffic source websites can have. So, if a site experiences a reduction in Google visibility, it may still continue to generate good traffic to and continue to prosper. Other sources of traffic include real ‘type-in’ traffic (when visitors type in a URL); social media traffic (ie from Facebook, Twitter, blogs and other); and affiliate traffic etc.
DaniWeb, which has been an ongoing sub-plot of the Panda storyline throughout the year, due to its victimization and full recovery, was hit again by the most recent update. In fact, Dani Horowitz, who runs the IT discussion community, is the one that tipped us that this was even going on.
Horowitz and her team have of course been doing some investigating themselves, and documenting this a bit in a Google support forum. In it, she writes:
So, everyone, thanks to DaniWeb’s handy dandy systems administrator, we have come to a conclusion. Our ‘time on site’ statistic decreased by 75% at 1 pm on August 11th, and has been holding steady at the reduced number, as a result of Google Analytics rolling out their new session management feature.
There have been MANY reports across the web of the bounce rate and time on site being inaccurate every since August 11th, especially when multiple 301 redirects are involved (which we use heavily).
As a result, we have been hit by Panda. Or so I gather.
Now, this is not confirmed, but could a Google Analytics change, and inaccurate data on Google’s part be responsible for sites losing over half of their traffic? If so, that’s not cool.
Google, who famously won’t reveal its secret recipe for search rankings or even list each of the factors without revealing the weight of each, has been historically vague about its use of Google Analytics metrics in search. Michael Gray recently wrote a post suggesting that you can almost guarantee that Google is using your Analytics data, but he mentions how Google always manages to sidestep questions about its use (or non-use) of data for bounce rate, exit rate, time on site, etc.
Another interesting side-story to the Panda saga is that Google-owned sites have done well (according to the Searchmetrics data). The timing of the most recent Panda update, which Searchmetrics counts YouTube and Android.com as major winners for, is interesting given recent Senate discussions about Google favoring its own content in search results. A Google spokesperson gave us the following statement on the matter:
“Our intent is to rank web search results in order to deliver the most relevant answers to users. Each change we make goes through a process of rigorous scientific testing, and if we don’t believe that a change will help users, we won’t launch the change. In particular, last week’s Panda change was a result of bringing more data into our algorithms.”
The Panda update has appeared to favor video content throughout its various iterations (and not just YouTube). I can tell you that video has some major SEO benefits regardless of Panda, and that it is also great for increasing time on site. If a user is watching a video on your page, they’re on the page for the duration of the video or at least until they lose interest (so use good video content).
Even Demand Media told us after they announced the eHow clean-up, that it wouldn’t much affect its YouTube strategy.
Update: Dani Horowitz tells us that DaniWeb has already recovered. More here.
Do you think Google is improving its search results with the Panda update? Let us know what you think in the comments.