SEO sites were blasted by Panda

[AlanBleiweiss]

I've never trusted compete.com data - every test I've done with their data has ranged from "wildly under-reported" to "wildly over-reported", with the occassional "lucky guestimate" thrown in.

[SteveOllington]

Yeah they do admit they estimate, based on their sampling. But I believe they mix panel with tool bar and ISP data, so I would have assumed their estimates would be kind of close but maybe not then lol.

Where do you use? I guess none can be trusted really can they?

[EGOL]

Thanks Steve,  I don't see an end-of-February drop for any of those sites in that data.

[TomCritchlow]

Yeah, all of these data providers are a little suspect. Especially compete!

[SteveOllington]

Yeah even if the numbers are way out with the different source, you'd expect to see some reflection with the trending. I wonder what's wrong with Trends then.

[SteveOllington]

"Especially Compete"... haha, worse then Alexa?

[AlanBleiweiss]

I'd like to say "you can only trust numbers if you have access to an individual site's anlaytics.  Except every analytics program is suspect.

Oddly because they're all suspect, I use data for finding trends in the data.  Which brings us full circle to Google Trends.  Apparently Google can't even spot accurate trends.

[EGOL]

Why did they all move the same direction on the same date - the panda date?

[bemcapaz]

I think maybe they do spot but its not in their best interest to just release it to the world for free

[SteveOllington]

Eureka! (Possibly... I think/hope).

Trends uses four sources right?

Sooooo, if you think about it, all it would take is for Trends to stop using just one of those sources and wham! There it is, the drop!

Now, I noticed people recently saying that GA had removed benchmarking, which will be replaced by something better later. Since GA benchmarking is part of the permission-al "opt in" data that makes up Trends, and the fact that it disappeared at the same sort of time... logically you could conclude that Trends suddenly stopped getting data from that source.

That would explain everything!!

I should have been a detective

[meshach]

Interesting thread! Has anyone got any third party to confirm? SEOmoz have anything to add?

[vaibhav45]

Our Site learnhub.com & Jumbotest.com saw 30% increase in Traffic after panda update.

[toeplly]

I am Managing SEO for Multiple Websites, One Health Website Traffic Goes Down Up to 60 % , Before Panda We are getting 3500 Daily Visits and Now It's 1700 Per Day, Help me Out, What Steps Should I Follow to Get my Traffic Back ..

[EGOL]

I have not heard of any sites that were hit by Panda being released from the problem.  If people were getting out they would be yelling YEEEEHHAAAA! in the forums.  So, until that happens you can only guess at what might be holding your site back.

I would focus for now on an overall improvement in content quality.   That is always a great investment.

[Websensejim]

Yeah, looks like Google Trends data is not to be trusted (like most other G products). I just looked at those graphs for some heavyweight trusted authority sites including the BBC and guess what - same big traffic drop  LOL

Steve O, I think you may be on to something.

[SteveOllington]

If it turns out to be right, I'm going to get myself a deerstalker hat, start smoking a pipe, and walk around calling everybody Watson.

[EGOL]

waiting for the new profile photo

[flowsimple]

Interesting thread. So now after the more recent Panda update, is that another dip I see in those same graphs There's gotta be something with wrong with the data. Most of the data I've seen from Google Trends and Compete is laughable.

So any further updates? How has Panda affected SEO websites? And why? Is it all the incestuous linking? Thinking out loud here... is it possible that Google has some way to benchmark what the mean/median/average links there are per industry and THEN it uses those to figure out what a websites link count is worth? (Know what I mean?) eg. SEO industry mean = 300 real links, so above 300 higher authority, but dentist websites mean = 60 real links, so different way to measure authority? Point is, links and their worth could also be classified and measured by industry.