Page size and keyword density

Posted in Links

Having a discussion at alt.internet.search-engines under the thread Re: Links and PR with Borek, and decided I’d put my answer here instead of on the NG. Won’t waste some half decent content this way :-)

Borek asked-

“Let’s see… Is there optimal size for a page SEO Wise?”

I replied-

For Google below 101KB and preferably more than a few paragraphs of text so your content includes derivative search phrases (based on the main phrase for the page) for maximum overall related SERPs, beyond that it doesn’t matter.
Borek replied (futire response from Borek are indented)-

Well, you are asking others to do some testing before posting their SEO ideas,

And that takes us to my latest response (only on this blog).

I don’t think I’ve said the above exactly? Pretty sure it was more along the lines that there should be less of the “well known SEO fact” type postings where a few regulars hold a belief, post about it a lot and before you know it, it’s a well known SEO fact. Best example is validation, where the evidence suggests validation per se is irrelevant.

but you are on your own doing the same mistake of posting things that can be easily falsified with simple test :)

You’ve made the mistake of assuming I haven’t run tests on this.

Size DOES matter – there is a bell curve with optimal size of the page. When comparing series of pages of identical keyword density, all other things being identical, pages of some particular size are ranked higher.

Assuming you did the tests, your tests are flawed since you are using keyword density as a variable which you later state there is no optimal keyword density (I’ve got similar results on keyword density tests BTW).

If there is no optimal keyword density how can you use it as a standardising factor in SEO tests?

You’ve made the assumption that as long as you keep the density the same you will only see the effects of increased/decreased page size (I’m assuming you mean amount of content rather than actual page size, since most code is ignored, it’s content that matters, feel free to test it :-)).

As you later rightly state the more you use a phrase (within reason) more likely it is you’ll get a SERP, however as the amount of content increases trying to keep a high keyword density results in spammy content!

I’ve found through various tests that-

high content + very high keyword density equals poor rankings

conversely-

low content + very high keyword density doesn’t equal poor rankings (doesn’t mean you’ll necessarily get good rankings though)

What I’ve deduced from this is if all you care about is one phrase, keep the page low on content and with a high keyword density, but if like me you want maximum targeted traffic ignore keyword density per se (don’t waste your time measuring it) instead make high content pages and include as many related phrases as possible on the page without going spammy, since one page can obtain more than one SERP.

As I’ve said many times here thousands of low traffic SERPs tend to result in more traffic than a few high traffic SERPs and the traffic is more stable as well. Drop a few places with a high traffic SERP and watch your traffic die over night, do the same with 50% of your low traffic SERPs and you’ll still have enough targeted traffic to survive while you gain more traffic.

What was a rather shocking discovery for me (reported earlier to the group) the same doesn’t hold for keyword density – there is no optimal KW density, the higher the better. I never tried above 50%, but it just doesn’t make sense.

I’ve got similar results and it does make sense if you look at this from the perspective of Google wanting high quality content, not spammy content in it’s database.

It’s easy to get a page to above 10% keyword density without going spammy if there isn’t much content, but try to do the same with a reasonable size page and keeping the keyword density up is impossible without going spammy.

So from Google’s perspective when it finds a page with a lot of content and very high density (relatively speaking) it can be reasonably confident it’s a spammy page. This can be incorporated into their search engine ranking algorithm without throwing out too many babies with the bath water :-)

You can see the whole thread from Google Groups at Links and PR.

If you wish to make a comment feel free to comment below.