Exactly Which Penalty(ies) Was I Hit By?

[Planet13]

Can you help me figure out why my site has lost about 45% of its traffic since about mid-April of this year?

Please check out this graph from google analytics:

[i43.tinypic.com...]

while it looks like Penguin 2.0 on May 22nd took its toll, it looks like traffic was already starting to slope down from around May 1st.

If anyone has any more insight into what possible penalties / algo updates might have affected my site, please share your opinions here.

(Irony here is that I have spent lots of time improving content, removing thin content, and I haven't done any link building since mid-2011)

[adder]

If you compare 3 sections in your timeline:

May 1 to May 9
May 9 to May 18 (another period of 9 days)
May 22 to May 30

Can you pull in some more detailed stats as to what got hit? Money terms? Long tail? Any particular group of keywords? Any particular section of the site got hit? etc...

Also, did you use any ranking trackers? Do you know which search terms dropped on 1st and 9th?

[Planet13]

Hi there, Adder, and thanks so much for taking a look at my post and the graphic:


In short, EVERYTHING is down about the same amount. all types of keywords, all types of landing pages - everything.

Unfortunately, I don't have any rank tracking services / software.

but I do have google analytics set up and can look at the TRAFFIC patterns FOR THE WHOLE SITE.

Down 8.81% : May 1, 2013 - May 9, 2013 compared to Apr 22, 2013 - Apr 30, 201

Down 18.56% : May 9, 2013 - May 18, 2013 compared to Apr 29, 2013 - May 8, 2013

Down 23.72% : May 22, 2013 - May 30, 2013 compared to May 13, 2013 - May 21, 2013

Pretty much ALL terms (money, informational, branded) were hit equally.

We have two separate sections of our site: One is an ecommerce, the other is a blog. The blog part (on same domain) links to our ecommerce pages.

Both the blog section and the ecommerce section declined equally.

We don't really rank for many SHORT tail keywords. Most of the keywords we rank for are three or four words long (or longer).

The few two-word keywords that we DO rank for are generally for our blog and one of the words is usually "meaning", so people are looking for info more than a product.

(I should mention that about 20% of the top 100 keywords to our site have the word "meaning" or a synonym for meanin in them - is that bad?)

Single-Word Keywords are down 35.29%

Two-word Keywords are down 32.28%

For branded keywords (that use the three words in our domain name) we were down over 52%, but on the other hand, we don't get much traffic via branded keywords anyway, so in real terms it was something like 24 searches compared to 50 searches for the previous period.

~~~~~~~

Really, I have been looking at my google analytics six ways from Sunday, and it just looks like EVERYTHING is down, pretty much equally. Whatever affliction I have seems to be sitewide, and statcounter seems to back this up as well.

[aristotle]

Well the Penguin demotion seems pretty clearcut. The rest of the lost traffic could be due to a combination of smaller effects, such as

(A) Your site could be among the many that have been gradually losing traffic as Google continues to push large well-known megasites and brands higher.

(B) When you make changes to a site, even if they are improvements, usually the initial effect is a temporary drop in rankings and traffic. If you keep making new changes, it could keep pushing you down again and again.

(C) If you got backlinks removed, or disavowed them, because of Penguin, the whole site is weakened due to the loss of incoming link juice.

So with so many potential factors involved, it could be hard to separate out their individual effects in a clear way.

[Planet13]

One other thing I should add.

I know one site that is scraping about 11 pages of our content.

Because they have a poor URL structure, those articles appear on several pages on THEIR site. Not just 11 pages, but about 65 pages of their site has my content on it.

I requested a DMCA with google to have THEIR pages delisted, and google said that I had not sufficiently proven to google that it was OUR content.

I am in the process of getting a new DMCA together to send to google, but since that person started using our content in March of this year, I wonder if that is having an effect?

[backdraft7]

You sure that's not simply seasonal decline? I see the same pattern from April to June.

[diberry]

I requested a DMCA with google to have THEIR pages delisted, and google said that I had not sufficiently proven to google that it was OUR content.

This is really odd. Google has never blinked at delisting scraped page I ask them to. I wonder what they look to for "proof" that you posted it first. Their caches? Something on the server? Please tell me merely sticking on a date, as I do, isn't enough, because anyone can doctor that, even in a CMS that does it automatically (I use Wordpress). I've accidentally scheduled posts for the time of the King John and Robin Hood before.

[adder]

Planet13, from what you said about scraped content, I suspect this might be some sort of Panda activity. Google stopped announcing Panda updates and said it would run as a hidden assassin alongside the main algo. But they surely must be tweaking the Panda "section" of the algo. So, it might as well be that the drop in your traffic was something to do with a Panda tweak.

(I should mention that about 20% of the top 100 keywords to our site have the word "meaning" or a synonym for meanin in them - is that bad?)

Do you think you might have over-optimised the pages that rank for, let's say "blue widgets meaning". Do you have the keyword "blue widgets meaning" in the url and the title and the H1 and stuffed within the content?

For branded keywords (that use the three words in our domain name) we were down over 52%

Right... can your domain name be described as EMD? i.e. your domain name is nicebluewidgets.com and you have anchor text "nice blue widgets" in your link profile? And can everyone understand that "nice blue widgets" is indeed your company name and not just some keyphrase or product name? The reason why I'm asking is that I've seen a lot of Penguinised sites lately and their branded terms are almost never affected. If a company "Widgetry Blues" is hit by Penguin, it is likely to retain the #1 position for the term "Widgetry Blues" and is even likely to display a full set of sitelinks.

[Planet13]

@ backdraft17:

There might be some seasonal variation playing a role in it.

However, when I compare year over year the period May 1st to May 31, 2013 is down 24% from the same period in 2012

If I compare the same period from 2011 to 2013, then 2013 is down 28%

If I look at the period just after penguin 2 (which struck on May 22nd of 2013) I am down 32% from 2012

If I compare from May 22nd 2013 through June 22nd, 2013 with the same period from 2011, I am down 41%.

[Planet13]

@ diberry and @ adder

You both raise good points.

I have to run out for the rest of the day, but could you please check back in this thread in about 8 hours? I want to respond to you but the wife is pulling me by the arm out the door.

[backdraft7]

@plate - I'm seeing the seasonal dip, PLUS the 30% down overall...meantime, G is raking in cash hand over fist. We're does it come from? It's our old long tail traffic diverted through their ad network pals.

[JS_Harris]

A few questions:

- Is this all web traffic or does the graph have image traffic included? Image search is a different beast, do you have a chart for each over the same period?

- How much traffic does the graph represent? If it scales from 150 visitors to 100 visitors the actual impact is different than if we're talking about a drop from 100,000 to 50,000.

- Which pages took a hit? Is the drop sitewide to all pages equally or are a handful of pages representative of more than 50% of the drop?

Graphs can be highly subjective without additional information.

[Planet13]

@ JS_Harris

"- How much traffic does the graph represent?"

Not HUGE traffic. I was getting about 750 unique visitors a day. Now it is down to around 450 per day. (I know, I need a better business model... I'm working on it.)

- Which pages took a hit? Is the drop sitewide to all pages equally or are a handful of pages representative of more than 50% of the drop?

For the most part, it was pretty much equal across ecommerce (all categories) and our blog (all categories and pages).

There were a couple of pages on our blog where it went from one landing a day to, maybe, 15 landings a day, or something like that. But for the most part, it was all equal drops in traffic.

"- Is this all web traffic or does the graph have image traffic included? Image search is a different beast, do you have a chart for each over the same period?"

I don't really know how to view image search traffic in google analytics, so I tried using this report that was created by Annie Cushing

[google.com...]

If I compare image search from the May 2013 to Apr 2013, then traffic from image search is down 17%

If I compare image search from May 2013 with image search from May 2012, it is down NEALY 80%!

However, I do not know how well that report actually tracks image search traffic. From what I understand, google frequently changes the referrer for image search, so it just might be that report that I got doesn't track it very well for this year...

[Planet13]

@ diberry:

"Google has never blinked at delisting scraped page I ask them to."

They are making life tough on me...

First there were a couple of ebay pages that used my content.

I filed a DMCA with google and a week later google said to look for ebay's own notification system to let ebay know that they have scraped content on ebay's site.

Next I filed a DMCA with google about the site I mentioned above that scaped 11 of my articles and was showing them on about 65 pages of his site.

Google responded like this:

We have received your DMCA notice. It is unclear to us whether or not you are the authorized copyright agent for the content in question. Only the copyright owner or an authorized representative can file a DMCA Infringement Notice on his/her behalf. Please note that you will be liable for damages (including costs and attorneys' fees) if you materia

lly misrepresent that a product or activity is infringing your copyrights.

I am not sure what that means, so I started a separate thread here:

[webmasterworld.com...]

But the gist of it is, I do not know exactly what google wants me to do to prove that I own the copyright to our own content :(

It shouldn't be hard for google to see that the content has been on our site for several years before the scraper's site even existed.

[Planet13]

@ adder:

"Do you think you might have over-optimised the pages that rank for, let's say "blue widgets meaning".

Anything is possible, but no, I have the word "meaning" in the title element of one page (our most popular landing page) and in the H1 element of that page, but it doesn't really appear frequently throughout the page or even throughout our site.

Oddly enough, I actually ADDED the word meaning to the title AFTER the page had already been ranking #1 for "widget meaning" to improve clickthrough rate, and because THAT WAS WHAT THE PAGE IS ABOUT (even though I left it out of the title originally for brevity).

"Right... can your domain name be described as EMD?"

It's really a partial match domain instead of an exact match domain.

I don't really know how to exemplify this without using keywords, so I will use some keywords that are NOT in my industry.

Suppose you had an ecommerce site that sold stuff for your kitchen (several different items).

And suppose a widget is an item that you use to slice bread, or slice meat, or slice things in general.

Our domain would be something like: The Widget Mountain.com

Even though we sell other kitchen stuff besides widgets.

But it wasn't just the pages that cover widgets that lost traffic. It was all the other pages that cover other kitchen stuff that lost traffic as well.

"And can everyone understand that "nice blue widgets" is indeed your company name and not just some keyphrase or product name?"

Well, apparently not. It turns out there is actually a PLACE called The Widget Mountain, but since it is an area of land, there isn't an official website for it.

And just to add to the confusion, on our site The Widget Mountain.com, we don't actually have any information about the physical place called the Widget Mountain.

"If a company "Widgetry Blues" is hit by Penguin, it is likely to retain the #1 position for the term "Widgetry Blues" and is even likely to display a full set of sitelinks."

If I do a search for:

The Widget Mountain

I am number two, no site links. The number one result is a yelp listing for Widget Mountain Restaurant in some backwater Florida town. (That restaurant is nowhere near the actual location of Widget Mountain).

If I do a search for:

the widget mountain.com (with the spaces)

It says:

Are you searching for thewidgetmountain.com ?

If I search for thewidgetmountain.com then I get the six pack. (Of course google, in it's infinite wisdom, has decided to list my xml site map as one of my six-pack results. Thanks google. Now I gotta fix that in WMT)

[adder]

Well, then it's not to do with the EMD. If it was an EMD penalty, you'd rank #1 for The Widget Mountain.

Actually, some webmasters have reported similar activity on or around May 9th.

Have you been hit by any of the previous Panda updates?

It would be interesting to take a date range from Oct 2011 to this day and match it with the Google penalty timeline, just like you did on the graph you attached.

Moz has got one: [moz.com...]

[Planet13]

Hi there, Adder:

Thanks so much for the suggestion to go back and look at my traffic and compare it with the SEO moz chart.

- March 23, 2012 (Panda 3.4): Drop of 32% in visits according to Google Analytics

Apr 19, 2012 (Panda 3.5) Another drop - this time about 22% in visits from google.

From May 1 2012 through Dec 14th, it is pretty smooth, but it is definitely a slight DOWNWARD angle. Dec 14th saw a 22% drop in traffic, but that is to be expected because of seasonality (most people finished up their online shopping around then).

Things were improving slightly (up about 6%) over the late January through Early March period of 2013, but then the graph looks like a saw tooth wave from about mid March until late April.

Then it looks like successive drops on:

May 1, 2013 (No known update)
May 8, 2013 (Phantom Update)
May 21st / 22nd (Penguin 2.0)

Wow! I did not realize that google HATED my site so much!

If I compare the Month of June, 2013 with my BEST month (which was January 2012), I am down 77% off my busiest month.

Wow... this has been an eye opener... I don't even know what to say...

[Planet13]

Yikes!

If I compare just the most recent period year over year it is depressing:

June 24 through July 24, 2012
Average 813 Uniques Per Day

June 24 through July 24, 2013
Average 474 Uniques Per Day

Well... so much for adding good content...

[netmeg]

Ok, step back from the content a minute and start looking at your user metrics. When people come to your site, what are they doing there? Are the numbers of pages per visit off too? Are the same topics / pages / products popular (relatively speaking) as last year? What about your bounce rate? Have done any usability testing on your site? Put in some heat map software (you can get it for as little as $9/month) to see where they're clicking and where they're not? Viewed it on multiple devices? Tightened up your site architecture, experimented with anchor text on buttons and links, analyzed your social shares?

Content is important, but it's not JUST content. If I recall, you were having issues where people were coming to your ecommerce site to view an information page, and then flitting off without buying anything. Is that still happening?

[JS_Harris]

I think netmeg is right on here, you need to figure out what's going on with your visitors to rule out usability issues and see if conversions can be improved. It's time for a new baseline that you can use to measure the impact of new changes over time which will, unfortunately, take months to complete. Make minor changes and watch for effect while allowing sufficient time for those changes to propagate.

Yada, yada, I know you understand what netmeg suggests but I wanted to add a note specifically about your index page. Take a look at your index page from the point of view of a first time visitor.

- Does it very clearly tell a visitor exactly what the site can do for them?
- Does it clearly tell a visitor why they want to bookmark the site and come back regularly?
- Does it signal authority? Quality?

I have seen some sites, mostly built on wordpress, that just display the latest 10 articles and these sites seem to have been impacted more harshly than others. If the index page content is ever changing there needs to be some very strong signals to answer the questions above. Google raters will be looking for these things on your index page every time they are asked to evaluate internal pages.

Lastly, it has to be asked, does this site have many affiliate offers? Sites with a lot of affiliate links don't seem to hold their ground as well with recent changes. They can do well but they require extra effort, the 'more' Google demands has to offset the affiliate link effect, especially if you have pages with a lot of affiliate links(or only affiliate content) on them.

Advertorials are under the gun right now, pages with only advertorial speak AND affiliate links are double in trouble.

If it's not an affiliate site, and you do sell product directly to customers, are your product descriptions unique or repeated from the manufacturer? Duplicate content is not only from scraped pages, it can be from repeated product descriptions.

[netmeg]

You can also get user testing where someone uses your site to perform some task (looking something up to buy, or finding products, or whatever) for as low as thirty or forty bucks too. They give you a screen cap of how they navigate your site, along with a sound track of their comments on what they're thinking. Very valuable, I dunno why everyone doesn't do it.

[Planet13]

Thank you both, netmeg and JS_Harris:

I will look into the usability issues.

Content is important, but it's not JUST content. If I recall, you were having issues where people were coming to your ecommerce site to view an information page, and then flitting off without buying anything. Is that still happening?

Yes, that is still happening. The vast majority of landings are on the info pages (in the blog). If I look specifically at the blog numbers in GA using content drilldown, they are like this:

Time on page: 02:56 (site average 01:30)
Bounce Rate: 78.34% (Site Avg: 74.99%)
% Exit: 69.66%(Site Avg: 53.30%)

If I compare Apr though June this year with 2012

Time on page is UP 7.5%
Bounce Rate is Down 2.2%
Bounce rate is DOWN

"Are the numbers of pages per visit off too?"

Pages per visit are DOWN 5.9% (overall from last year).
Down 14.5% from same time 2011

Avg. Visit Duration is UP 2.5% from 2012
Down 7.6% from 2011

UPDATE:

Are the same topics / pages / products popular (relatively speaking) as last year?

Unfortunately not:

It looks like year over year the BLOG (information) pages lost only 6% of landing page visits.

However, my ecommerce pages are DOWN 74% as landing pages!

I am just realizing this now. The only positive for my ecommerce pages is that the pages per visit average went up 6%.

But otherwise, it was the ecommerce pages that got slaughtered.

I will try to dig into this some more. I am apparently worse at google analytics than I thought...