Forum Moderators: open
How could INK know the "theme" of a site unless they spider more than 1 or 2 pages (assuming they aren't paid). Most of the sites I have that aren't paid never get spidered beyond the root and I light off fireworks when they pick up the odd page here and there.
I don't know the answer to your question rogerd but I noticed the same thing too.
I think that's the problem. For older Best of Web sites, Ink may have spidered thoroughly and "know" what the theme is. For a new site that pays for a relatively small number of pages, there may be insufficient data to determine a theme. (There's some discussion of BOW at [webmasterworld.com...] .) The high-ranking sites that I'm seeing at the top of the list appear to be older, well-established large sites that could benefit from either site theme, link pop analysis, or both.
Assuming that I'm on the right track, how does one combat this? Paying for hundreds of pages to be included almost certainly wouldn't have a favorable ROI, even if decent Ink rankings were the result. Can on-page optimization beat off-page criteria in Ink?
Great question - have you determined that he pages in question are paid or older. (easily done at MSN by looking at the URL and looking for an INK redirect)
From what I can tell the paid pages are like little Islands. They are only using their own resources to rank so link pop has nothing (or very little) to do with them.
Are the keywords competitive? Did you check the full html for the KWs including combo boxes, etc?
Can you give an example URL and the keywords in question so I can look at it - here or in my Sticky Mail?
-s-
There are pages in the INK db where it appears INK sees them as the source for information even though there is no reason for them to rank.
Search Honda Accord here [169.207.238.189...]
There is no reason for Honda.com to be number one - but a lot of reason for the number two page to be #1
Cash??? or Directory style listing??? who knows???
-s-
I also understood from them that not all of the linking pages would necessarily show up in the public Inktomi data base... that they might be in the Web Map but not in the database of pages returned on searches.
When PFI began and it seemed that Inktomi was not following links... or at least not including in the public database the pages it found by follwing links... I wondered aloud on some forums (no longer having access to the Inktomi rep) whether in fact Inktomi had dropped link relevancy.
It may well be, though, that Inktomi is using its Web Map for link relevancy, but only showing new paid pages. Lots of posts here have reported that pages are being spidered but not included; this could be why. It could also be, though, that Honda.com has been in the index for a long, long time.
This creates a problematic situation for new sites - if spidering and indexing are down, and off-page criteria are still important, what does one do to combat older, entrenched sites? Pay for including pages on other sites that are linked to you (if they aren't in Ink already)? Would these links even be counted, one wonders? Could get costly, too...
"if spidering and indexing are down, and off-page criteria are still important, what does one do to combat older, entrenched sites?"
Its a double edged sword for us:
1) Really happy that some of our old pages are immovable and hold their own with no work needed
2) Displacing competitors old pages with new ones which seem to fit everything that is required except that they havent been their for some time is very difficult.
Interesting dilemna Ink's in:
They hold onto the pages in B.O.W in order to keep the quality of their data up there and resellable to search portals, at the same time, they are aggressively promoting the PFP model.
Got the feeling that one morning, I will wake up and my old pages and the traffic that they generate will be gone!
I think one resolution might be for Ink to strive for the most complete database using free spidering for everyone (like Google does), at least every few months, but offer PFI pages very frequent spidering and a small rankings boost. To compete with the big sites, one would still have to employ good site design and optimization, still need external links, etc. At least, though, the rankings wouldn't be as static as they are now.
Overall, PFI is a dangerous step onto a slippery slope, IMO. If you include everyone for free, why should anyone pay? OTOH, if you add only those who pay, the quality of the search results deteriorates rapidly. Finding the middle ground that delivers quality results to searchers but encourages enough people to pay is indeed tricky.
IC provides an XML feed for these sites.
From Search Engine Watch:
"One feature with Index Connect is that Inktomi can take in an XML feed. This means that instead of crawling a web site, you'd feed it a database that would list each URL, its title, meta description and keywords tags, plus body abstract or other information."
This may be the reason why a Honda site tops the charts for no apparent reason. The spider is not crawling the site.
Original SEW article:
AV appears to describe a similar setup with their Trusted Feed program.
Just round up a few clients/affiliates for a couple of hundred pages each and submit them all in bulk. No up-front $$, just .25 per click and a #1 ranking. (Charge your customer .35 when you get the bill)
Beats the average per click rates with Overture
.
Also, I've got the unexpectedly strong performance of one of my own pages for a keyword that is important for the site as a whole (both site theme and Looksmart listings), but not for that page. I still have sense that something is being overlayed on the page analysis... Anyone else see this, or is this an anomaly?
It's almost like a sliding scale...the more competitive a term, the higher the keyword densities.
