Forum Moderators: Robert Charlton & goodroi
Not sure how this will paste from excel, but here are my data points:
primarytotal%IBLp/IBL
site01 7,770 25,400 30.60% 1,970 3.94
site02 107,000 813,000 13.20% 760 140.79
site03 833 8,510 9.80% 30,400 0.03
site04 3,800 44,600 8.50% 7,410 0.51
site05 3,410 42,800 8.00% 421 8.1
site06 916 11,600 7.90% 700 1.31
site07 19,800 251,000 7.90% 3,090 6.41
site08 1,350 18,400 7.30% 211 6.4
site09 1,880 30,800 6.10% 343 5.48
site10 1,040 33,300 3.10% 362 2.87
okay, looks like those columns won't align, so, you'll have to stick with this one... the first data column 'primary' is the # of pages in the primary index, the second column is the total # of pages in the index, the third column is the % of primary indexed pages to total indexed pages, the fourth column is the number of inbound links showing in yahoo and the fifth column is the number of primary pages per inbound link. note that i sorted these results descending by percent of primary indexed pages to total indexed pages - they are not in the serp order (but they are the top ten of the serp).
a couple of comments first - the sites with the four highest totals of IBL's have bought lots of links. some of these are kind of obvious, you know, sitewide footer links etc. all of the sites have been up for at least a few years, remember, i am considering this 2 word keyword combo to be moderately competitive.
ok, now for some analysis. the first thing that strikes me is how low the primary percentage is for these established sites. site01 performs the best in terms of getting the highest % of pages in the primary index. it is not the site with the highest tbpr (obviously google has degraded many of its bought links). generally, though, the sample sites aren't doing very well. random keyword checking shows extremely poor results for the supplemental (well, this was the old term, right?) pages - for example - on a typical greyed out page that is not in the primary index (supp) page titled 'fuzzy green metallic widgets' is not even in the serp unless you use "'s - then it comes up - this was typical across all of the sites - the grey bar is pretty much death - even on obscure and specific longtail keywords.
so why do some of these sites perform better than others? ok, i don't know. going back to the old supplemental discussions, the reasons that a page could be supplemental (well, these are a few...) were 1. insufficient page rank 2. dup content 3. poor anchors/linking structure (related to #1) 4. not enough original content (related to #2). since all of the sites have similar homepage tbpr and have reasonably similar (and good) internal linking architecture, i am not thinking 1 and 3 are the answer.
here are some questions that i would like to pose. how do on-page factors affect whether a longtail page will be supp? do big sites keep their % of primary pages - in other words, if i know i can get 9% of my total pages in the primary index, why don't i just make a million page site? any other aspects that we should be focusing on here? anybody else care to share a dataset?
I'll have a stab at your questions :)
how do on-page factors affect whether a longtail page will be supp?
On-page factors are highly significant, and seemingly small factors seem to be reason enough for a page to lose the ability to pass benefit or value and/or be supplemental.
do big sites keep their % of primary pages - in other words, if i know i can get 9% of my total pages in the primary index, why don't i just make a million page site
I don't think it's a percentage game, but a question of how much content of desirable quality a single site can support. It's also common for sites to dramatically increase the level of low quality pages via the addition of a high quality page. A product often comes with (unpopulated) review pages, recommend a friend pages and the like. Some sites create a link to every comment posted.
Can you explain how you determined which pages are supplementally indexed and which aren't? If you are using one of the "punctuation hacks" then this is useful information to collect, but is not necessarily a reflection of:
- A URL's ability to rank
- A URL's ability to pass benefit to other pages
i used the old /* to separate out the primary and sup indexed pages - i realize it is a crude tool. interestingly, though, i checked 10 grey bar pages from each site on their obvious keyword and almost none were even in the serp at all. i have seen grey bar pages in serp before, but not recently and i couldn't come up with any examples on my site.
since these sites are all ranking pretty well on competitive terms, they are pretty similarly seo'd (except for the sites that paid for links...). of course there are differences, but, by and large, all of the sites have the keyword in the title, h1's, keyword in url etc. - this makes it harder to discern small differences in on-page content and structure that could affect the grey bar curse. i am increasingly thinking this is it, though, so i'll have another run through the sample sites tomorrow looking carefully at on-page stuff.
From my experience, supplemental pages can result from a single factor (large or small) but also a combination of factors. So, it can be difficult to isolate specific reasons, depending on the data you're looking at. Whenever I've taken samples, there is not usually a reliable degree of correlation between a single factor in isolation and a page's performance. It's often a judgement call as to what factor to address - where experience and knowledge of SEO play at least a big a part as stats can.
Make sure you also look at internal/external link structure, and if possible links right down to the page level, since deficiencies there seem to be one of the most common reasons for degraded pages. And by the same token, high quality links (internal or external) to supplemental pages are enough to cover most vices ;)
since these are all ecomm sites, i don't think natural inbounds will come so easily. it does seem though, that the sites with more inbounds outperform on the % of pages in the primary index yardstick. so maybe it is more off-page (as i think marcia's question implies).
i think the big question is how google values those links that some of the sites went out and got - they are generally of lower value, but they can be fresher and, after all, they are links.
one thing that is interesting about the data - 6 of the 10 sites have between 7 and 9 percent pages in the primary index - should we benchmark this for large ecomm sites?
