Google Updates and SERP Changes - February 2013

[petehall]

< continued from [webmasterworld.com...]

Not just location screwing things up... all web browsers are now using your history to distort results, even with Web History switched off. Only way to check the real results now is Chrome Incognito.

That and the location issues are really screwing things up. If I type in weather, G gives me weather for Windmill Hill, UK... I don't even know where that is. I looked and its 200 miles away at the other end of the UK!

[edited by: tedster at 2:08 pm (utc) on Feb 1, 2013]

[petehall]

20% increase on one of my sites. Not seen much obvious movement though. Could just have been a busy day !

[ethought]

Interesting that a site that used to toggle with mine for the #1 slot is now at #456 with mine at #457.

That is very interesting..

[backdraft7]

@husky Pup - those are my product conversions, not adsense ads. I do not run a pure MFA site, but I do run one adsense tower. What's funny is that my adsense income used to outpace my adwords spend nearly two to one. Now, two years after larry's arrival, and no real changes on my part, it's dead even.

Yesterday I had to cancel Alert Site services and now considering downgrading my server to a shared account. My income has now dropped over 50%. There's just no fighting the well funded (multi million dollar) sites who have moved in and taken up entire pages.

[reos]

busy days are the best. we're still seeing a drop in traffic that started in the beginning of the month (feb 3). there was an update on jan 31; anybody else seeing that lag and effects? a small but crucial (read, highly competitive) group of keywords dropped out of the first page and on to the second and we'll have to target these with off-site campaigns pretty heavily to make up the gap. We weren't penalized in the sense that we didn't receive a message (<-- though I'm not sure if that's an accurate guiding light), but we were definitely impacted by the algo and I want to stay away from page content optimization right now

[Lorel]

@reos

I experienced a slow drop on Jan 31 and eventually lost about 1/2 my traffic (which is about 1/2 of what it was 2 years ago) and it's still slowly dropping.

There was a similar drop Starting Dec 2 then traffic slid for a month until Dec 29 but that didn't affect main keywords so I assumed it was due to the holidays (I don't sell products - only services).

Then traffic picked up again until Jan 31 (reaching the previous high peak) when it started dropping again slowly but this time it has affected main keywords. I had already deoptimized all those pages months ago but I just did it again and updating pages I haven't touched in a couple years also. I have several pages that still rank highly and bring a lot of traffic but not income generating pages.

This isn't the first rollercoaster ride this site has experienced in the last 2 years although it has been better than last year. I am assuming it's because I hadn't updated enough pages last time.

I don't see a panda update around Jan 31. Have you seen one?

BTW, I have had some success fixing pages where the title and H1 tag were using the same keywords and also repeated in first paragraph.

[reos]

@lorel: we're seeing a light uptick now but this puts us on par with last year, not above it. we're rolling out a side-wide fix on naming protocol for the meta, so if you've been seeing the h1/title changing be helpful then that bodes well for us.

[Wilburforce]

I have had some success fixing pages where the title and H1 tag were using the same keywords and also repeated in first paragraph

@Lorel, what effect did this have?

I have quite a lot of pages that use the same words in Title and h1, and in general it hasn't appeared to do any harm: my use of Title and h1 is generally "what the page is about".

e.g. (w3.org): "If the document is basically stand-alone, for example Things to See and Do in Geneva, the top-level heading is probably the same as the title."

It doesn't reflect well on Google if this attracts a penalty.

[reos]

what's the similarity ratio? there's a difference between descriptive meta and stuffed meta. are the h1 and titles different from each other? when i came on they were the same, so that would get hit fairly quickly

[Wilburforce]

If my page is about widget polishing, both Title and h1 will generally be Widget Polishing.

I don't stuff Title or h1 at all ever. Obviously having Widget Thingumy Whatsit You-know Device as Title and h1 is not sending Google the same message as Things to See and Do in Geneva.

[reos]

and you didn't see a drop, even with them matching?

[Wilburforce]

and you didn't see a drop, even with them matching?

A few pages with matching Title and h1 have lost/regained/lost position in the last year. Many more, also with matching Title and h1, have not dropped at all.

I do not attribute the drops in position to matching Title and h1.

[SEOmy]

As a curiosity, how many of you that have seen dramatic drops have adword accounts that Google know you've used in the past to promote the site that dropped? A bit of a conspiracy theory I know - but one I think is interesting.

[tedster]

I do not attribute the drops in position to matching Title and h1

You're smart not to do that, too. Many commercial Content Management Systems will create that matching by default. The web is full of such pages that are totally innocent, so IMO Google cannot use this is a spam signal.

[Lorel]

I may be wrong about duplicate Title vs H1 tags causing a problem however I've seen drastic improvement when making sure they are not exact duplicates of each other.

Also, when the title and H1 are exactly the same I either change the words around or use a synonym instead, but still maintain the same meaning.

BTW, I haven't changed the titles - I only change the H1 as I don't want a delay in ranking due to title change.

@Tedster Re Commercial CMS programs creating matching titles and H1 by default and Google not having a problem with it. I'm very concerned about leaning on what some software program does by default and hoping Google is sympathetic towards that method, i.e., the default that uses numbers in file names instead of keywords for instance. When Google is trying to get rid of duplicates it seems to me having a slightly different title/H1 can't hurt.

I don't use a CMS program. I write code by hand. Maybe Google can tell the difference, i.e., it sees a deliberate attempt to duplicate title/H1 in my case.

[GreenDog18]

I can also second the the fact that my rankings increased once I changed up my H1's to differ from my title. Might have been something else but I didn't make any other changes.

[tedster]

I've also seen rankings improve after changing H1 and title so they aren't an exact match. And every time we can, I ask clients to customize their CMS so they can make them different, and train their content team to make intelligent choices.

It's just that I don't see Google inflicting a true penalty on a site just because there's an exact match between H1 and title.

[gouri]

For those that saw improvements after varying the title tag and h1 tag, were they exact match before you made the change, or were they pretty similar and you made it even more different?

[TheOptimizationIdiot]

It's just that I don't see Google inflicting a true penalty on a site just because there's an exact match between H1 and title.

Do you think it could be something more along the lines of a combination of things and the H1 is just the first thing people change when adjusting other aspects of the phrase based spam detection could have the same impact?

[0133] In the first section (first m entries), the following relevance attributes are stored for each document entry in the posting list of a given phrase: [0134] 1. The document relevance score (e.g., page rank); [0135] 2. Total number of occurrences of the phrase in the document; [0136] 3. A rank ordered list of up to 10,000 anchor documents that also contain the phrase and which point to this document, and for each anchor document its relevance score (e.g., page rank), and the anchor text itself; and [0137] 4. The position of each phrase occurrence, and for each occurrence, a set of flags indicating whether the occurrence is a title, bold, a heading, in a URL, in the body, in a sidebar, in a footer, in an advertisement, capitalized, or in some other type of HTML markup.

[0042] M(p): Number of interesting instances of the possible phrase. An instance of a possible phrase is "interesting" where the possible phrase is distinguished from neighboring content in the document by grammatical or format markers, for example by being in boldface, or underline, or as anchor text in a hyperlink, or in quotation marks. These (and other) distinguishing appearances are indicated by various HTML markup language tags and grammatical markers. These statistics are maintained for a phrase when it is placed on the good phrase list 208.

[appft1.uspto.gov...]

[tedster]

Do you think it could be something more along the lines of a combination of things and the H1 is just the first thing people change when adjusting other aspects of the phrase based spam detection could have the same impact?

Exactly!

[Zivush]

good phrase list 208

Trying as much a possible to have more phrases on the good list is one of the best SEO practices.
I don't know how many read the patent document.
I think it is a must.

[Wilburforce]

Do you think it could be something more along the lines of a combination of things

I am certain that it isn't one factor in isolation. It is conceivable that there is a "tipping point", and changing the h1 tag *might* make a difference on its own, whether or not it is the same as Title.

However, having both the same is widespread, generally conforms with W3C advice (see my previous post), and makes obvious sense where both correspond with the subject matter of the page.

Pracitces that are more likeley to incur a penalty might be:

1. Where they do not correspond with the subject matter of the page;
2. Where they are used for keyword stuffing.

I am as sure as I can be, however, that identity on its own is not a penalty factor. I have several pages still at #1 for the phrase in Title and h1.

[MikeNoLastName]

This specificity is all so absurd! G get a life! They are acting like a spoiled little baby. They only want from us what they are too lazy to create themselves, and punish by crying because they can't have it served exactly how they want it on a silver platter. If all the relatives would just stop paying attention to the wah-wah little G-baby in the corner of the room he might stop crying.

BTW, after reading it, what the "208" patent referred to says to me is, that if your web page is unique, creative and "not almost EXACTLY like all the other pages out there, then you're SEO spam". It IS Silicon Valley mentality after all so let's all sing it out loud "Little boxes on the hillside, Little boxes all the same.... And they're all made out of ticky-tacky..."

[edited by: MikeNoLastName at 11:30 am (utc) on Feb 27, 2013]

[Jez123]

This specificity is all so absurd! G get a life! They are acting like a spoiled little baby. They only want from us what they are too lazy to create themselves, and punish by crying because they can't have it served exactly how they want it on a silver platter. If all the relatives would just stop paying attention to the wah-wah little G-baby in the corner of the room he might stop crying.

That would make a pretty funny google doodle :)

[neuroseoservice]

Which methodology is favorable to retrieve SERP ranking on google?

[Wilburforce]

I don't know how many read the patent document.
I think it is a must.

If that is what is now being implemented, it actually throws some light on this mess.

It has at least one fundamental flaw: if the existing data set contains imperfect pages, then predictions based on it are weighted by those imperfections.

Presumably, that is why good pages are sinking and bad pages are rising: there are more bad pages than good pages, so any phrase-distribution "standard" is weighted in favour of bad pages.

Presumably, also, that is why thinner pages are rising: a good page with a lot of content about a subject will contain a higher number of related phrases, while averaging on the basis of the whole population of pages predicts a lower number. There are relatively few good pages with a lot of content, so they are more likely to fall at the extremes of distribution (and attract a penalty).

To paraphrase MikeNoLastName, it is Conform to the Norm or die.

[TheOptimizationIdiot]

Bold added

4. The method of claim 1, wherein identifying the document as a spam document, further comprises: responsive to the actual number of related phrases present in the document for at least one phrase exceeding the expected number of related phrases by at least a multiple of a standard deviation of the expected number of related phrases, identifying the document as a spam document.

Based on the preceding I doubt it's exactly "conform to the norm" since you have to be over by a multiple of the expected standard deviation to be considered spam. The quotes from earlier were to point out what is actually taken into account which is much more than most realize.

I think Zivush gives some good advice to people about reading the whole thing and I'd add reading all of the phrase based scoring patents can be enlightening.

[scottsonline]

What I'm seeing in the past week that had been gone for a year:

1. For a popular product in our niche:
- the #1 site has "this is a demo store" above the page. It is a Jon-functioning test site.
- #2 is the manufucaturer
- #3 thru #7 are seller1.manufacturer.com, seller2.manufacturer.com and so on template sites provided by the manufacturer for resellers. Content on 2 thru 7 is identical down to the punctuation.
- #8+ actual reseller sites with custom content.

This is happening over a broad number of searches. Google has dramatically boosted the relevance of what are essentially affiliate sites if they are hosted as subdomains on the trusted site. Lol. Terrible experience for shoppers and browsers.

[Wilburforce]

at least a multiple of a standard deviation

could mean 1.1 x SD, or most of the data.

it could not realistically mean >3 x SD (or effectively all the data).

However, if most pages on widget tinkering have the phrase once or twice, a comprehensively informative page (widget tinkering history, widget tinkering in early belgian convents, modern widget tinkering methods...) would almost inevitably fall foul of it.

[TheOptimizationIdiot]

Eh, read the whole thing. It's way more complicated than you're making it sound and it's not exactly like you're saying about the number of phrases, cause they use one phrase to predict the likelihood of the presence of another (one or more), which means amount of text/content on the page would have an influence on what phrases are expected and the expected rate of the phrase being repeated.

Saying a longer page that repeats a phrase or related phrase more than a static number of times it's usually repeated on a shorter page is not what they're doing or trying to do, for exactly the reason you state.

Another way of saying why they aren't and wouldn't do it that way is: It would be silly, overly simplistic and likely not patentable.

Example of Predictability or Expectancy Change by a Phrase:
The phrase "Comprehensive Information About Widgets" or something similar could (would in my opinion) impact the expectancy of the phrases you're stating would cause any longer page to tank, cause the expectancy or predictability of other related phrases would likely be impacted so the longer page wouldn't tank due to the predictability of those phrases in conjunction with a more "inclusive phrase about widgets" being present on the page and those phrases themselves would impact the expectancy or predictability of other related phrases on the page.

[Wilburforce]

read the whole thing.

Yes, read the whole thing.

Paragraph 4 is *one* of the proposed methods to implement paragraph 1, and has the flaw I have outlined.

The whole proces is flawed, and making it complicated does not improve it.

It would, if implemented, explain the current mess. If lots of people start changing their pages on the basis of it (they have already), the distribution of phrases will change, and pages that were not previously more than 2 standard deviations from the mean will become so.

As a result the spread will keep diminishing (it is killing off outliers on the distribution curve). It biases the web's evolution away from diversity.

Conform to the Norm, or die.