Forum Moderators: open
First of all, the site is to be read by people, and they expect a logical arrangement of data, topics, subject matter and so on. Of necessity some of those pages may have to be quite long.
So if we start chopping those long pages into Google sized bites, what value are we bringing to the viewer? They'll have to follow a multitude of links all over the site when one single page will give them a better experience.
Menu structures start to balloon out with the attendant problems of managing the site navigation.
I'd like to think that the content, be it on long pages or not, will always outweigh a contrived preference when it comes to offering viewers the best sites.
More importantly it would be very difficult to proove.
Just looking at statistics of page size on SERPS says little as you have to correlate that to the average sizes and distribution thereof in the whole Google index and you would have to discount the fact that smaller pages are probably those seen as most "usable" and most "seo'd".
If you have extremely good content people will link to your page.
If you split that content over several pages, you will have less focussed and less amount of links compared to the original one page biggy and therefore rank less high than with the split up.
On the otherhand in less competitive areas, splitting up will mean more (motivated) internal links, more differentiated specialised links and more titles.
Ranking-wise I would keep it to a limited amount of pages if the field is competitive and the title can cover the whole topic of the page.
Lots of articles have been written about the necessity of writing in another way for Internet. Reading on a screen is something quite differnet from reading a peace of paper. I myself am in the proces of splitting my longer documents into smaller peaces. A complicated proces, but I am certain that it has to be done.
I've been mindful of the several posts, Brett included I think, which suggest Google likes (ranks higher?) smaller pages. We also know that content is king. If I am a provider of a big informational site, trying to meet both objectives may cause me some agravation.First of all, the site is to be read by people, and they expect a logical arrangement of data, topics, subject matter and so on. Of necessity some of those pages may have to be quite long.
Not necessarily true. The vast majority of online users don't read web pages -- they scan web pages for information of interest. Many print to read, bookmark for reference, save, copy to a word processor, etc.
So if we start chopping those long pages into Google sized bites, what value are we bringing to the viewer? They'll have to follow a multitude of links all over the site when one single page will give them a better experience.
One word - "Scan-ability".
Menu structures start to balloon out with the attendant problems of managing the site navigation.I'd like to think that the content, be it on long pages or not, will always outweigh a contrived preference when it comes to offering viewers the best sites.
Short 1 topic paragraph pages are best, the use of bullets helps alot: Also adding downloadable PDF's allows much flexibility, print friendly pages - right length, less ink, etc. and your long documents but only for those who want them.
A recent resource in some of these topics:
Software Usability Research Lab
Dept of Psychology
Wichita State University
Updated November 5, 2002
Criteria for optimal web design Designing for Usability [psychology.wichita.edu].
[edited by: fathom at 10:04 am (utc) on Dec. 22, 2002]
agree - the general public.
Yet if you're working in a scientific field, you tend to love those complete multi print page pdf's. As fathom says, quick scan, then print.
I've been mindful of the several posts, Brett included I think, which suggest Google likes (ranks higher?) smaller pages.
In most cases the "size" issue being discussed is kb per page.
It's usually design elements, not on page text, which contributes to the inflated page sizes which can cause problems.
As for page length in usability, testing with your own audience is the only sure guide because there are always exceptions to every rule.
I have split-run tests which show a higher conversion ratio for a long single page sales letter vs. the same information broken down into multiple pages.
Some data here [searchengineworld.com...]
Could it be that the proliferation of very small page sizes that rank higher is simply due to higher keyword density, as opposed to Google's "preferring" small pages?
Even though writing the pages for the scanners would make more of them happy, I write my pages for the readers. It is the readers that build my reputation and they are also the ones that give me the high PR links. The scanners can just go to epinions and get their short blurbs from other scanners.
Some of those large content pages with a PR2 rank higher than smaller specialized pages on other sites with PR5. And I do no conscious SEO on any of my content pages.
In other words, I don't believe that google really shows much preference for smaller pages.
Yes, i think you are right!
In fact my 250k site map ends up on the front page for hundreds of keyphrases
In other words, I don't believe that google really shows much preference for smaller pages.
hmmm... don't even need to look to say your chosen keyphrases are "not that competitive".
I would guess less than 500k competitive pages.
One look at any "single" word in Google and comparing the file size results across any market/industry would show that your belief is flawed.
Try competiting at 5 - 10 million+ pages... not a single 250K page to be found when you need one! ;)
An example: web design - 6 million competitive pages
33K is the largest page in the top 30.
If we reduce the competiveness 4.6 million for: internet marketing
A 55K page got to #1. In the top 30 it is the largest file size.
You're dead wrong if you believe file size matters not.
[edited by: fathom at 2:23 am (utc) on Dec. 23, 2002]
I would be wanting to present a relevant and trouble free list of results to searchers, and I would definitely want to penalize any pages that would be slow to load.
There is a very clear message in that graph NFFC linked us to in msg 6. Those figures are indeed quite incredible. Over 70% of the best ranking sites have pages less than 20k.
In my field it is dam near impossible to create a page with any sort of viewer appeal and get anywhere near 20k.
Which brings me back to the question in the opening post. If content sites are delivered on big pages, then maybe content alone may not be such a rankings winner as some folks think?
Oh yes... Merry Xmas to you all and your loved ones.
I'm not sure that the graph isn't open to interpretation. Perhaps it also reflects that most people create smallish pages.
I assume that the size is that of html and text only, and images are not included. So some of those pages could considerably bigger when viewed by a real browser.
However I am still going with the idea that small=good. :)
The SEW graph is - from what I understand - a spidering of the ODP at that time.
(it does not say when)
I am probably missing something, but can anyone elaborate on what the graph is trying to show, other than the average page size of ODP listed pages?
- Could one conclude that most ODP listed pages are index pages?
- Could one conclude that most external inbound links go to index pages?
- Could one conclude that index pages mostly rank best for competitive SERP's?
- Could one conclude that most index pages are not 6 screen pages?
Just as a line of thought:
If page size is to have a relevant statistical influence on ranking results, it should be an overriding factor in one of Google's many ranking variables.
But page size is an "on-page" factor and in my world on-page factors only play roles in uncompetitive fields, beyond that "in-site" factors (internal anchortexts and their net PR) are more important, and beyond that "off-site" factors (external anchortexts and their net PR) are more important.
If page size would have any statistical ranking importance, I could beat a competitor with more external inbound links just by playing with my page size.
Being a developer of Shockwave applets which tends to be near full screen size, much of the content of the page is "off-page".
With only a modest amount of text below the applets on the page -- these pages compete at the top against million of others.
Bear in mind that many other optimal characteristics help to produce this -- the 3 - 5K file size rings true as well.
The best is #1 against 4.5 million competing page on a two word keyphrase, 53 text words on page, 1 DMOZ listing PR6 (and Google directory PR7) , 27 external link PR4 - PR6 (noted so far) about 50 internal links, and 3,386 bytes.
In the same context and with the same aiding characteristics -- a 250K file would likely not be in the top 100.
If page size would have any statistical ranking importance, I could beat a competitor with more external inbound links just by playing with my page size.
Sure, if all other things were equal. But how often are other things equal, with so many variables in a Web page and 100 or so elements in Google's algorithm?
Now for two comments on other opinions that have been shared in this thread:
1) The primary benefit of shorter pages is likely to be more focused content. Let's say you're writing a 1,000-word treatise on muffins. By breaking the article into separate pages on its various subtopics (in this case, different types of muffins), you increase the chances that you'll rank high in Google when a user searches on "bran muffins," "banana walnut muffins," or "blueberry muffins" rather than just "muffins." So yes, you'll have better SERP placement and more referrals with shorter pages, but because of what's on each of those pages--not because of page length.
2) PDF files may be appropriate for online copies of printed user manuals, illustrated product brochures, white papers, and other documents that are intended mostly for printing, but they're a nuisance to view on a computer screen, and you can't assume that the average user will have Acrobat Reader or even know what a PDF file is. For Web pages that will be read mostly online, stick with HTML.
It is not the number of results that matters, it is the competitiveness of the results. And I will admit that my site map only shows up well on uncompetitive searches, though I am front page on a million+ result search.
My point was that my site map should be in the crapper on the results, given the things that we know or have guessed about various on page or off page factors. The URL in no way represents the search terms. The only links to it are from my navigation bar, with the deluxe text of something like "all our reports". The same for the title. The page contained 1300 links and was 266k in length. The anchor text of the links on the page were the URLs of the reports. It has a PR3.
It is one of the few times that I have done anything to try and manipulate a SE, I was trying to get google to crawl all my content, I never expected to have the page rank on it's own. I put it up while a monthly crawl was in progress, and as of a couple of weeks ago, it is no longer linked from my navigation.
This page that breaks all sorts of rules is somehow competitive with pages that are small, with equal or better PR, and significantly better content. Would it be competitive on a search for "web design"? I seriously doubt it. But my doubts are not based on the number of pages or the size of my page, but on the effort that goes into those pages.
While Google may give some credit to smaller pages, none of the arguments so far have convinced me of that. Several of the other posts here have pointed out other reasons that it might appear that smaller pages are better, and I think those have to be considered before deciding that google does actually measure such things.
But like I said, it was a last minute attempt to get google to crawl all the way through my navigation to get to the real content. Google had already started crawling and I tossed together a sitemap based on a DB search, in the hopes of getting google to go to as many of my content pages as possible. I didn't have time to break it up into seperate pages.
As it turns out, I don't think it was necessary, it crawled all my content that month anyway, including all the pages linked after the 101k point. I do not consider such a large page to be a good thing, and I have since removed the link to it. The page is still there, but that is only because I refuse to remove content and serve 404s to people who have a valid url.
Simply put, that sitemap sucked! It was an embarassment and a waste of bandwidth. I never wanted to get good results with it. But even with all links removed from it, it is still the 12th most popular page on my site just from SE referals.
It is not the number of results that matters, it is the competitiveness of the results.
I share your first point, adding that is most cases a phrase with only 3,000 competitive pages, isn't likely to be too competitive itself, and likely not receiving much traffic for that particular query.
And I will admit that my site map only shows up well on uncompetitive searches, though I am front page on a million+ result search.
Your second point was my point exactly. You can get a 1mb page to #1 on a "10 word" quoted query - but it doesn't prove or disprove anything.
Here is a fact - you can only go so far will any single technique in optimal design; for number of pages, internal links, internal PageRank development and manipulation, text, keyword density, Page Titles, Meta data, element attributes, outbound links, reciprocal links, inbound links, Natural PageRank, etc., before there is an appreciable drop in adding to ranking potential.
Most of these are generally hidden from view as to the precise point where "no more potential" can be achieved and all are effected by each change to any as well as changes in the Google archive by "Google" as well as everyone else listed. This make any appreciable evidence difficult to assess.
However, Google places file sizes on SERP's -- look at the most competitive phrases in your particular markets, and use the visable results to make your conclusions as to how much weight is associated with file size.
I will admit that my site map only shows up well on uncompetitive searches
Your site map's, file size and ranked position on a "not so competitive search term" taints your ability to develop conclusive evidence to prove or disprove anything to yourself and to others.
Please note:First post: In other words, I don't believe that google really shows much preference for smaller pages.Second post: While Google may give some credit to smaller pages, none of the arguments so far have convinced me of that. Several of the other posts here have pointed out other reasons that it might appear that smaller pages are better, and I think those have to be considered before deciding that google does actually measure such things.Not attempting to convince you of anything. The evidence is in the results..., and the research of "much results" versus a single page. Just pointing out the holes in your conclusion, so others can make there own conclusions
