Big changes in Google content scoring

[ipopbb]

Very good info about the "keyword matches"

For example, if i was promoting "Wheels":

Would it be beneficial to rename all images on my page that is targeting the term "wheels" to wheels-1.jpg , wheels-2.jpg,etc? Or even make blank images(that match the bg), named follow the same naming scheme.And would google penalize my site for excessive use of this method?

I can't answer that...

"What will Google do if..." ... heck if I know.

Call it Tagging files or Media Slugs or filename slugs or semantic filenames... CNN does it, Amazon does it, Target does it, NexTag does it, ... I do it.

 

[MisterGemini]

3. keyword density (if you don't have 50+ keyword matches per page (full source... check it out... even HTML comments help now) ... )

I can confirm by my own analysis of the Mayday update that this factor appear to be so, especially when it comes to product based sites.

 

[ipopbb]

I can confirm by my own analysis of the Mayday update that this factor appear to be so, especially when it comes to product based sites.

All my production sites are moving to the 50+ matches requirement... was a bitch building the system to make it tunable without it crossing the line into keyword stuffing... I followed a lot of examples I mention frequently here on BHW, Amazon, Target, SEO.com, Addicting Games.com, PinkRibbonShop.com

* see my thread comments on Spliffs for more details. (Its like an easter egg hunt!)

 

[target]

Prior to Caffeine/Mayday this tuning trick made my clients about $3M US... It doesn't appear to matter as much anymore since Google's algorithm is soo different and only has 4 significant factors

1. domain match
2. dmoz & yahoo links
3. keyword density (if you don't have 50+ keyword matches per page (full source... check it out... even HTML comments help now) ... )
4. page rank (Its back in a big way... It looks shaky because it is a secondary sort but it pumps traffic from google... A new twist on it is that if you want to maintain High PR you need to link to page... PR has been tweaked to favor home pages and index pages, another totally exploitable pattern)

I had been reading the whole thread from a while. Any observations on how the new G algo behaving other than the above 4?

 

[abluka]

Good informations, so i know that

 

[ipopbb]

I had been reading the whole thread from a while. Any observations on how the new G algo behaving other than the above 4?

Funny you should ask! See my latest thread!

 

[schnibitz]

When you say 50+ keys/page does that also include variations of the keyword, or does that mean 50+ of the exact keyword?

 

[theentry]

Maybe an update on this one? Looks like really good info

 

[robotrock]

ipop that is a genius plan having a magic keyword catchall for factor fluctuations on 250 sites and a beast to test on! i've got both of those, but not up to 250 yet lol i'd love to hear more about the methodology in general

 

[revvente]

Google is a moving target.

 

[MisterGemini]

Google is a moving target.

Which is why most prefer to use a shotgun blast method in trying to hit them.

 

[unknownassailant]

Sure is good information but for me most of the acronyms you list are useless as I dont k ow what they stand for.

Now get to work and spell out every single acronym and what it means and what it stands for.

Better yet why dont the forum members detail them one by one. You can start at no. 2 as I and I am sure most people know what the title is.

Why don't you look them up?

 

[vil4u]

Hey ipopbb...update please.

 

[arasferra]

H1's ranking in your tests does not make sense to me at all, H1 is a very important tag, featured in Google's SEO documents as well, after <title>, its considered to be one of the most powerful ones.

 

[polocolto]

if you don't have 50+ keyword matches per page

If your main keyword is long tail like "pretty big red car" you have to repeat it 50+ times as well?

I totally agree with you when you target a single keyword like car or games then you have to repeat it dozens of times. Some sites repeat it 100 times, but i'm unsure about repeating 3 or 4 long tail keywords 50+ times.

 

[SEOpie]

ipopbb we would like to her more from you

 

[zenithdoom]

H1's ranking in your tests does not make sense to me at all, H1 is a very important tag, featured in Google's SEO documents as well, after <title>, its considered to be one of the most powerful ones.

it's precisely it's mentioned by Google and so many SEO experts that it's abused frequently, fell from grace and treated not as important by the Google algo.

 

[ipopbb]

Hey ipopbb...update please.

This experiment is offline currently... New google only showing a sampling results even when using site: for the experiment pages... I think reflects changes that deployed around the time Instant launched.

I think it is because the test case pages are so small. My other sites don't get this sampling effect. Once I figure out the sampling behavior's trigger and thresholds I plan to set up a new experiment soon that will conclusively prove whether one kind of HTML tag helps content rankings over another and whether or not tags behave like a currency and what the exchange rates are... a.k.a. does 3.6 H3 tags = 1 H1 tag (assuming exact match in tag content)... the new report will rank all tags on how many H1 tags they represent. If there is no statistical bias then all tags should equate to 1 H1 in value. Of course I'll share the test case source again once I have the experiment method worked out. Feel free to roll your own experiments if you don't want to wait. Its not idea theft if I'm publicly sharing. I encourage people to try these kinds of things because I tend to be slow and methodical and usually pretty busy too. Just please share some of your findings!

ted

 

[ipopbb]

If your main keyword is long tail like "pretty big red car" you have to repeat it 50+ times as well?

I totally agree with you when you target a single keyword like car or games then you have to repeat it dozens of times. Some sites repeat it 100 times, but i'm unsure about repeating 3 or 4 long tail keywords 50+ times.

You have to think in terms of search types.

People don't usually use quoted search terms... at least the bulk of the search traffic isn't doing that. So if your goal is targeting traffic you need to think in a broad match search context. That means your page source (not necessarily visible HTML) should have

50+ pretty
50+ red
50+ big
50+ car

You could repeat the exact phrase 50 times or just use all the critical words a lot. Its why you should use full URLs in your source if you have keywords in them. Check out seo.com... they use full URLs which lets them use SEO in the source a great deal more without violating the webmaster guidelines in a blatant way. See my posts on spliffs for more.

If you want to know how much keyword stuffing is needed to be competitive then look at the top 10 results for your keywords and count the broad match usages. For some of my keywords it goes as high as 500-1000 usages.

Thats the new google. It's the worst google ever in terms of quality of results. But the good news is that it is pretty exploitable if you can figure out the 4 or 5 factors that seem to matter the most and implement them in concert.


Ted

 

[ipopbb]

Thanx Bro, please will you complete this sentence...does this website solely rely on SEO or more like PPC for traffic?

No healthy business relies solely on one channel. Whatever channel I focus on ends up being the growth leader. Email, PPC, Social, SEO, PR, conversion tuning, viral & sticky, ... they are all cash cows and when you use multiples you get a gestalt effect that is usually greater than the sum of the parts.