Can't identify on-page OR off-page factors that make page rank well.

[Onders]

We've got a site with a few sections where we review widgets. A lot of the links are to affiliate sites.

All the sections are pretty much the same and contain a couple of pages or so, but one main page with a lot of text. I've been comparing a couple of the sections to see if I can isolate some of the differences.

The first thing to mention is that one page / section does incredibly well for very competitive terms in its niche. The other section doesn't do very well in the other niche - which is also very competitive.

Now - we update the pages very frequently to input the talk about the latest widgets, so with regards to keyword density, metas etc, it's always changing and you would think that overall rankings would massively fluctuate because of this - but they don't. Both pages have lots of text on them by the way.

So - on page factors are constantly changing, so I can't quite see why one continues to rank highly, the other one not highly.

With regards to internal linking - pretty much the same as each other. It's been done carefully, but not over optimised, but the internal links relevant and on-topic.

Ok - so external links in to the page. Again, I've looked at these, and on the whole it's very similar. Neither page has many links from other sites, and the ones that there are, are from small directories. Anyway, anchor texts etc are pretty similar (relatively!) so I don't think this can play such a big factor.

Ok - so what does this leave me with? The real differences that I can pinpoint are:

a) One of the sections is update much more regularly than the other. The one that ranks highly has text changes every couple of days, the one that doesn't rank only has text changes once a month or so.

b) Visitor retention - my stats programme (a paid one) tells us that people stay longer and there are less bounces / people leaving than the badly ranking page

c) Page age - the badly ranking page disappeared from the index about 15 months ago (it was doing better before that), and it took me 5 weeks to get it indexed again.

So - how important can the above factors be? Page updating is an interesting point, but if a page isn't updated often, does that mean it may be of less use to a visitor? That doesn't make sense....

Visitor retention - ok, we go back into the debate about usefulness. A user can visit a certain page, find the info they want and click away in seconds on some types of sites. Visitor retention in my eyes can't be that important (controversial!)

Page age - again, a new page can be just as useful as one that's been there for 8 years. Can this really be that important?

Anyway - apologies for venting my frustrations, but I hope it raises some interesting points.. Do people agree with these potential factors - and to what extent?

Thanks!

[edited by: Robert_Charlton at 8:20 am (utc) on Dec. 31, 2008]

[tedster]

These all seem like possible, but relatively lightweight, factors to me. Possibly another factor may be Google's different handling of different keywords, particularly with regard to their statistical studies of user intention. Also, note if there is any semantic ambiguity possible in the keywords.

Also note that Google reps have very recently said that bounce rates are too noisy a signal for them to use as a ranking factor in the organic SERPs - as well as being a signal that could easily be faked.

[Robert Charlton]

How carefully have you examined the competition for the target phrases?

If all else were equal (something which I feel is very hard to gauge), is it possible that the phrases addressed by the badly ranking page are more competitive than those of the better ranking page? The fact that the page disappeared for a while might suggest this.

[Onders]

Thanks for your replies and thoughts Tedster and Robert. When you talk about "statistical studies of user intention" do you effectively mean whether the keyword is a buying keyphrase or an information gathering keyphrase? The main keyphrases for both pages are similar in that they are formed with a certain demographic (the same for both pages) followed by the widget they are targeting. Does that make sense?

Interestingly the better ranking page is potentially slightly less keyword heavy - they've both been written for the user and we've been careful to ensure there is no unnatural volume of keyphrases.

With the better ranking page, there are are more variations and related keyphrases, which we utilise quite well. The worse ranking page has a very limited amount of "surrounding / related keyphrases". This may well make a difference.

Another reason I think about page age is because a competitor ranks number 1 for the particular keyphrase, with a very flimsy and weak page. Their backlinks are nothing to write home about and there is very little content on their page. Interestingly, their number 1 spot has been cemented for a while, and very big direct merchants take the next few places.

You could well be right about the competition - with the badly ranking page targeting more competitive phrases. It doesn't quite explain why the current badly ranking page did rank well before it disappeared from the index (a redesign at fault for that - even though we were extremely careful to keep all content and internal links the same... but I accidently removed a homepage link for a while)

Anyway - any more thoughts much appreciated!