|Panda/Penguin and the use of hN headings / new thoughts|
| 8:55 pm on Sep 27, 2013 (gmt 0)|
Today I decided to see my new G ranking for a keyword which I was on first page for about 10+ years. I am now hovering around the 400 mark - considerably worse than it has been the last few months not that the difference was noticeable after the original drop.
Late 2011 and early 2012 is when the site was first affected then the emd update literally killed it. What I noticed though was a competitor who also had page 1 rankings for years was right beside me around 400... they had not been affected until recently. I do not know the exact date, but it had to be within the last month or so.
They are not an emd, but I noticed one similarity with our sites... headings on the index page. Both of us are using headings in correct format with one h1, a few h2s, some h3s, and what not. When comparing to sites who have not been hit I realized they use heading VERY scarcely on their index.
Is it possible that headings (in my case I had 13 total... 1-h1, 9-h2, 3-h3 which made totally logically sense) be considered 'spamming' by the algo? Visually and semantically these were used correctly and made sense for the reader, but...
Secondly, is it possible G is using a 'control' of high ranking sites and using that as a base to compare? As I said, all other sites use them sparingly or even just one h1 on the index and nothing else and have retained their rankings all these years. So because the 'control group' hardly uses them my 13 uses along with the other site who fell off the map as well were seen as gaming the system? This is an interesting theory to me and if true then there would be no 'true' seo as every niche would be compared to the rest of the sites in that niche. This would make sense as it appears to be one of Google's goals (remove seo), however, I can see lots of problems with its usage if true. It would explain why 'bad' sites are ranking and 'good' sites are no longer if the control group is all 'bad' sites... and vice versa of course.
Anyone else come across anything like this or notice similarities? Is the 400 level a new penalty similar to the 300? I honestly never kept up much with that as I never needed to worry about it.
Just another thought that popped into my head. I am testing changes and will report back, but for the moment this gives me some hope as it means my emd is not the problem perse, but rather some new trigger I hit.
If anyone uses the moz addon in FF my site and the other stick out like a sore thumb at the 400 level as our pa and da are very high compared to everything around us has essentially nothing. Not saying page authority and domain authority are the magic solution to rank, but it does provide a reference compared to other sites.
| 1:37 am on Sep 28, 2013 (gmt 0)|
I would guess it has more to do with the words in the <hN>'s than anything else. I haven't seen your site, so I can't say this is the issue for sure, but back-in-the-day repeating the topic of the page [keywords] in the headings was a good idea -- Today, from what I've seen it's not such a great plan.
If you think about reading a book that has a Title and 9 Headings [chapters] how often does the main topic need to be repeated through the book for readers to understand what it's about? In-my-experience, it doesn't, yet it seems to me many site owners think it's better to be "repetitive" rather than "expansive" in headings they have on any given page.
As I said, I can't say this is the issue with your site, but it's something I think could be an issue for some.
| 1:56 am on Sep 28, 2013 (gmt 0)|
Yeah, that is not the issue at all.
| 2:02 am on Sep 28, 2013 (gmt 0)|
Then I'd look other places, because I doubt it's the use of headings personally -- Of course some of it depends on how the document structure is interpreted by a machine reading system, but they're getting fairly good at "getting" what markup means what, so I'd say the chances of it being the headings are slim.
One thing to keep in mind about headings though is every time there's a new heading it's implicitly the beginning of a new <section> of a document, so there are cases where I think it's possible an algo can get "confused" when an <h1> opens, then an <h2> follows, then an <h3> [which should be a subsection of the <h2>] follows, but the <h3> is interpreted as another "main section" of the document, rather than a subsection of a section of the document -- Two or 3 repetitions of those type of "interpretation errors" down the page, imo, could possibly be an issue and it would definitely be easy to miss.
Hopefully that makes a bit of sense, but I'd still guess the use headings themselves aren't the issue personally -- I'd think it's more along the lines of structure or phrases associated with the headings if they're even part of the issue.
| 2:21 am on Sep 28, 2013 (gmt 0)|
The sparing use of H tags is something I have noticed in the top 5 spots across my niche with many different keyword targets. I am doing a few test changes to see if there is an effect.
| 2:52 am on Sep 28, 2013 (gmt 0)|
JD - thanks for your responses, but I am well aware of how headings are used and should be used correctly. My 'theories' may seem strange and they are as I have been scratching my head for some time (over a year) trying to figure out what Google does not like (unless of course EMD is the only reason for this). I've been around awhile so these crazy ideas I have are just that, but that is all I really have to go by. People I know personally have looked at this site and do the same thing - they have the slightest clue as everything is pretty much by the book and definitely moreso that those still at the top.
I get its hard for someone to understand without knowing all the details of the situation, but that kind of sums it up.
taber - I would be interested to hear about your tests so please keep me updated. I realize my OP sounds 'out there', but this is kind of the first correlation I have been able to put together other than just saying my emd got screwed.
| 3:12 am on Sep 28, 2013 (gmt 0)|
|My 'theories' may seem strange and they are as I have been scratching my head for some time (over a year) trying to figure out what Google does not like... |
Who hasn't been in that spot with the "oddities" of rankings these days? I definitely understand.
Let me move to a question: How much content do you have under each heading [<section>] relative to those ranking above you with fewer headings? In-other-words, do they have less headings with more depth of content under each or is your page just way longer with a relatively equal amount of content under each heading as those who are ranking?
| 3:17 am on Sep 28, 2013 (gmt 0)|
If the EMD update "killed it" then your site must have almost zero links and the few it has must be very low quality. The EMD update basically made it so that you could not get to the top of page 1 by buying an EMD domain and submitting it to a dozen directories. In other words, it targeted EMD sites that were ONLY ranking because of the EMD boost. If your EMD site has even one or two dozen quality backlinks, you can still get it to the top of page one.
I'd suggest you go back to your analytics and look at the EXACT date that your site was "killed" and correlate it to some Google update... Panda #20 rolled out two days before EMD and likely took a couple of days to settle down. Unless your site has only a few low quality links, I'd be willing to bet your problems throughout 2011 and 2012 have more to do with Panda than EMD.
I doubt very seriously your problems have anything to do with headers... and more likely has to do with content... low quality, duplicate, or "thin" content... all of which are targeted by Panda.
Of course, since we can't "see" your site, we're all just guessing based on limited data.
| 4:24 am on Sep 28, 2013 (gmt 0)|
Definitely does not have a 'few' links... it is going on 14 years old now.
|low quality, duplicate, or "thin" content |
Nope... and that is not coming from just me, but many friends/acquaintances who have personally looked at it.
Like I said, scratching my head as well as others. I appreciate everyone chiming in, but I have heard and read all of this over and over again for quite some time.
I'd actually like to get back to my original questions. No offense to the site or anyone, but as everything is confidential in the public forum this is how topics get somewhat off topic.
- Minus 400 penalty?
- Anyone else notice headings coming into play... scarcely used in top of their niche?
- Thoughts that every niche has its own 'control group' of ranking factors they compete against? This would eliminate the same de facto rules for every site assuming they were spread across different niches and competing for different terms. Google wants to get rid of seo and this would be a way to do that in the sense that each niche would have different ranking factors/values than the next causing different results.
| 12:57 pm on Sep 28, 2013 (gmt 0)|
It would be really sad if Google is using such a minor thing as this to trigger a major rankings demotion. But in truth I wouldn't be surprised when you consider how bad Google's search results have gotten over the past few years.
| 1:35 pm on Sep 28, 2013 (gmt 0)|
|t would be really sad if Google is using such a minor thing as this to trigger a major rankings demotion. |
Yes, it would be, but is there any hard evidence to suggest that Google is doing that? Speculation doesn't make it so.
I'd be leery of overreacting to blind guesswork--especially since such goofiness on Google's part (if it did exist) probably would be caught and corrected, not set in stone.
| 1:55 pm on Sep 28, 2013 (gmt 0)|
Honestly, I think you're spinning your wheels focusing on headers as the root of your problems. This sounds like something much bigger than a single HTML element.
Before you can treat an illness, you have to first know what it is. Otherwise you're just stabbing in the dark. And thinking that "headers" caused your traffic to die over two years... and eventually took out a competitor just doesn't seem at all likely to me.
I would take a really close look at your analytics and note EVERY date throughout 2011 and 2012 when you noticed any type of big, sustained drop in traffic (if such drops actually occured... I'm assuming based on your OP that it didn't just slowly dwindle over two years).
Correlate those dates to the various updates from Google (you can get a list if you search for something like "Google Updates Timeline"). There is one on this site somewhere... Moz and SEL both have easy to use chronological lists of updates as well.
If you can correlate the dates where you saw big traffic losses to specific updates that should lead you in the right direction for finding a cure. If for example, they correlate to Panda updates then I would be looking at your content. While you might not think content is an issue, it still could be.
Also, have you done a detailed backlink analysis? Google was filtering and penalizing a lot of companies around 2010 and 2011 for links (though quietly... this was before they became more tranparent). If you know which URL was ranking for the keyword phrase in question that is now around 400, I would look at the link text used in the backlinks to that URL (primarily external links, but also internal).
I would also look at the external link placements themselves... are they on bad neighborhood sites, are they sitewides like footer or blog roll links, are they along side links to totally unrelated sites which could cause them to be misinterpretted as paid links, are there any paid links, are there a lot of low quality links (like unnatural links that you or some SEO team built from directories, article submission sites, forums, blogs, etc.), etc.
And finally, have you filed a Reconsideration Request for the changes you've made to date to attempt to recover? If not then filing one should get you back a response that might at least let you know if there is some type of manual action on your site (regardless of whether the RR results in them lifting it or not).
| 1:57 pm on Sep 28, 2013 (gmt 0)|
There is no "hard evidence" there is only correlative evidence coming to the surface. More and more people are noticing the lack of h tags in ranking pages.
This could be caused by a tweak to the h tag algo
This could be caused by too many matching keywords in h tags causing a flag
This could be caused by a content requirement between h tags
Correlation is not causation. Google has gone to war with anything SEO. So it would not surprise me if there has been a algo tweak that targets what was standard SEO practice concerning on page SEO.
edit due to sleeping in English class in High School.
| 2:07 pm on Sep 28, 2013 (gmt 0)|
|Yes, it would be, but is there any hard evidence to suggest that Google is doing that? Speculation doesn't make it so. |
As for "speculation", this whole thread was started to discuss this possibility. As for "hard evidence", the OP had noticed a correlation, but how can anyone expect total proof in a case like this.. Unfortunately, a few people here tend to make a lot of posts without ever actually saying much. I generally try to skip past their posts but sometimes forget
| 8:13 pm on Sep 28, 2013 (gmt 0)|
Geez, it gets tougher and tougher to actually get a conversation and questions/opinions going on here that actually sticks to the topic or is understood correctly these days :).
I'm not focusing my problems on anything I just happened to notice a correlation and figured I would see if anyone else noticed anything similar. I'm not new to all of this so please give myself and others I have spoken to on a personal level for their opinion some credit - we actually know some stuff lol. The consensus is and has been 'in the air' and nobody really has a clue what has happened here. That certainly includes the basics such as how a header should be used, checking backlink profile, and everything else. Reconsideration requests are absolutely pointless unless you have a manual penalty applied to your site. Please lets keep the 'do this/check this basics' out of this conversation.
In this day and age with Google it is my opinion that topics like these, when the conversation works, is the only way we can better understand the underlying nature of how the algo is working. I noticed a correlation (for the first time mind you) and am now putting in some time to see if the correlation has bearing - that is all. If we work together on things like this it will help everyone as a whole.
There is no hard evidence obviously, but when I try a few things I will be happy to report back for those that care. I was just looking for some 'have you noticed this' input and 'make you think' conversation on the topic.
| 10:41 pm on Sep 28, 2013 (gmt 0)|
I've seen no correlation that "headers" 1, 2, or 3 make a difference. I have seen, however, that sites older than ten years (one of mine's nearly 20 years, ie, before G) have recently been hit hard... even with constant updates on site to standards. Keeping this reply on target: Headers are not the problem, never have been. Things like keyword abuse, whether text or headers... well that's what Panda, etc. have all been about. And sans a look at the G algo in use (never happen!), it's all a shot in the dark.
That said, create your docs correctly, with headers, text, etc. Though it sucks big time, traffic still comes from -400. Other choice is to quit... and for every quit out there, G and the players gain just a little.
| 1:16 pm on Sep 30, 2013 (gmt 0)|
I have noticed the lack of any <Hn> tags of the sites on page one for a little while now. Your dealing with a flawed system so you can't assume what is logical.
| 1:40 pm on Sep 30, 2013 (gmt 0)|
If you were going to describe how to spot spam to a 12 year old, what would you tell them to look for? <H1>, <H2> tags and the term inside them. The repeated use of a word or terms on a page etc. Now with that, look at your page and see if that logic would trigger anything. Now every one of us will say yeah but common sense says they should know that isn't an attempt to spam. Right there is the issue. This algorithm doesn't have that common sense to realize those little factors.
| 3:26 pm on Sep 30, 2013 (gmt 0)|
|I have seen, however, that sites older than ten years (one of mine's nearly 20 years, ie, before G) have recently been hit hard... even with constant updates on site to standards. |
It could be that the algorithm used to give points for domain age and no longer does. That's as likely an explanation as a Google jihad against "H" tags.