Forum Moderators: open
We have sacrificed much of our UI, dynamic content, page functionality, linking patterns, text placement, and limited our use of tables to please the Google.
Here are some of the changes:
- killed all .asp pages and went strictly to .html
- use text links rather than drop down boxes or forms
- removing nested tables that had been used to improve our page's appearance and structure
- removed javascript functionality and other header content
- ignored META tags for content and description
- cut down on headers, footers and other duplicate content that would be useful to have on each page
Anyone want to confess anything similar you did for your pages or a client's pages?
Here are some of the changes:
- killed all .asp pages and went strictly to .html
I think the issue of dynamic pages is often misunderstood. The problem is not quite the fact the pages are dynamic, but rather dynamic pages often take too long to serve, probably because they are working on some heavy task like querying a database.
Therefore, Google gives more time between crawling sites that respond slowly as Google want to avoid overloading crawled servers that may not be capable of handling bursts of crawling requests. This makes Google crawl much less pages than it would if the site responded quickly.
If you converted you dynamic pages into static pages, that will help but it was probably not necessary, if you used for instance content caching systems or script execution accelerators. I can give you may examples of sites that server dynamic content that are fully indexed by Google, including many that use extensions of Web scripting languages and even? and parameters in the URL.
I do not use ASP myself to recommend anything. I use PHP and I can recommend this class for caching arbitrary content [freshmeat.net] (partial or full pages) and this PHP script accelerator [turck-mmcache.sourceforge.net].
- use text links rather than drop down boxes or forms
Yes, Google follows plain links, not forms.
- removing nested tables that had been used to improve our page's appearance and structure
?!?! I do not understand this one. Google indexes visible text. If you said you eliminated DIV and SPAN that Javascript in them, I would understand, but tables, I do not see the possible reasoning.
- removed javascript functionality and other header content
- ignored META tags for content and description
Yes, Javascript in links is a big No-No. I think header content does not matter because it is not visible to the user, so I do not think Google cares about it at all.
- cut down on headers, footers and other duplicate content that would be useful to have on each page
If it was useful for the user, why did you remove it?
This suggests therefore that the first column of the table is read first, second column next etc. To the user the information on the screen seems to have equal preference (Above the fold if you will) however to the bot the first 30 or so words it encounters in the first column of the table will consider the primary words on the page and therefore the importance could end up on the navigation or image or ads rather than the actual content on the page. Doesn't the placement of text on the page matter to the bots?
footers/headers
- the headers and footers included on each page or many sites could suggest highly duplicated content on linked pages
You're looking for the old "table trick" which puts the content ahead of left-side navigation:
<table border="1" cellspacing="0" cellpadding="8">
<tr valign="top">
<td width="160">Our site:</td>
<td rowspan="2" width="480">Neque porro quisquam est, qui do<b>lorem ipsum</b> quia dolor sit amet, consectetur, adipisci velit, sed quia non numquam eius modi tempora incidunt ut labore et dolore magnam aliquam quaerat voluptatem. Ut enim ad minima veniam, quis nostrum exercitationem ullam corporis suscipit laboriosam, nisi ut aliquid ex ea commodi consequatur? Quis autem vel eum iure reprehenderit qui in ea voluptate velit esse...</td>
</tr>
<tr valign="top">
<td>
<a href="http://www.webmasterworld.com/">Lorem ipsum</a>
<br>Cicero, 45 B.C.
<br><a href="http://www.webmasterworld.com/">Lorem ipsum</a>
<br>Cicero, 45 B.C.
</td>
</tr>
</table>
To answer your main question, I changed one title to eliminate a repeated-as-plural keyword, since it's not longer needed now that Google is using word-stemming. I also added a few more pages of useful content to another site that needed some help. Other than that, no changes except to update pages which are updated regularly anyway. The recent "Brandy" adjustments seem to have returned the sites to their former good spots.
I agree with mlemos: If your footers are useful, then by all means leave them in!
Jim
- killed all .asp pages and went strictly to .html
The best way to get dynamic content spidered is to create static links to that content. The problem wih dynamic content in terms of spidering is that it's ... well ... dynamic. It's not necessarily fixed or stable at a single URL that can always be relied on to be THERE. However, Google (and most search engines) are of the attitude that if you are willing to make a static link to this dynamic content, you are indicating that the content is and will be THERE, so they'll take the chance of spidering it.
There are, of course, exceptions to this. You can make all the static links you want to dynamic content that includes a session ID, for example, and it will not likely get spidered.
- use text links rather than drop down boxes or forms
All search engines are plenty capable of following links in drop-down menus. No need to stop using them.
- removing nested tables that had been used to improve our page's appearance and structure
You're into overkill here. There are tens of thousands of web sites ranking very highly in Google that have tables nested inside tables nested inside tables. CSS-based layout is still very much the cutting edge when you consider the millions of people everyday who are uploading web pages to a web server.
- removed javascript functionality and other header content
I'd say more overkill here, too. Only thing I'd be concerned about is if your javascript code is so bloated that it starts to impact content density/keyword density, etc. But there are plenty of highly ranked sites using javascript.
- ignored META tags for content and description
The Description meta is very important, and if well-written, can be one of your greatest assets in gaining search engine traffic. Most search engines will display the Description as part of their SERPs. Google will, too, unless it finds the actual search query in the content of your web site -- if so, it'll show that rather than the Description meta.
- cut down on headers, footers and other duplicate content that would be useful to have on each page
More overkill. Duplicate content penalties apply to the actual "meat" of your web site, not to the matching dinnerware on which every page is served. :) Common headers and footers are a basic element of user-friendly site design, and shouldn't be discarded for fear of losing search engine rankings. Again, look at all the web sites ranking highly in G that use common headers and footers across the entire site.
Hope this helps. As G itself says, focus on what your users want, not what the search engines want. Don't let fear of G lead you to make decisions that hurt your site's usability.
In the last post the comment regarding the "meat" of the page is interesting. Any ideas or comments on how G determines what is "meat" and what is "dinnerware"?
Golden rule is to keep URL's short?product=123 and never use?id=123 as the id is very disliked by Google.
As for removing Meta Tags well I think thats a mistake, because Google still reads Meta Description and Inktomi reads Meta Description and maybe the keywords as well.
Clean html code is not only good practice but does help spiders find the content easily. CSS will help and these days I try to have only 1 table then the rest in CSS so I can still keep a well aligned web site.
No duplicate pages and plenty of internal text links and a few outbound links.
To be honest I haven't changed my site at all just added a page a day and improved my management of the site so it takes me less time to update and maintain.
My experience with dynamic content, ASP.NET serving .aspx docs:
- New Y! loves them
- G spiders, indexes 10% - 20% but takes months from spidering, indexing, showing URL only, then showing title and finally showing the whole thing, my URLs are not too complex though but do have id=?
- IIS or ISAP rewrite of URLs works excellently server-side (serving html-static-looking URLs upon request) as long as you do not have too many, needs to be supported by sitemaps and other text links leading to them / serving them, so it won't work with 200,000 pages
- mirroring, rewriting into into static actual phyiscally existing pages: has worked best, G eats them up like chocolate pudding, indexes fast, title, description immediately, serving PR and ranking higher as well
If you diversify between agents like serving dynamic pages to Y! (wouldn't change it in the world as it is 90% more than the static versions) and serving up rewritten statics to G and other SEs, make sure to allow, disallow with your robots file or you might run the risk to be penelized for duplicate content.
My sites that are currently dropped use, in one case, HTML 4 with all layout as <div> not a table used anywhere. This site passes w3 validation.
The other one is XHTML all divs for layout and passes w3 validation.
My site which is #1 for everything uses very complex tables produced by a visual design package. This does not pass w3 validation becuase of occasional non standard tags.
I was starting to think that Google did not like CSS divs layout.
Best wishes
Sid
In the last post the comment regarding the "meat" of the page is interesting. Any ideas or comments on how G determines what is "meat" and what is "dinnerware"?
It'd be one of the easiest things to figure out -- the meat, in almost all cases, is the section of the site that has the highest text density ... or maybe I should rephrase that as the highest text-to-code ratio. It often begins with an H1 or a larger font size and then goes into regular paragraph structure. A lot of the table/layout code, or javascript code, or whatever will be up in your header. A lot of the navigation links will up in the header, or in a sidebar, or down in the footer. But you don't typically find that stuff in the meat of a page.
Search engine programmers are quite smart. They can sort this out and you don't get hurt by using the same header, footer, whatever across your site. Why would they penalize you for doing something that's user-friendly? :)
No, I have not changed anything on any of my many sites since the last two updates. Google can't keep this crap they call an ago. It's all about increasing the ad word revenue. They will have to turn back sooner or later.
As a result of not changing my sites, they are now in the top 10 for 90% of my keywords on Yahoo. My Google results sink, but Yahoo traffic and conversions have trippled my sales.
It's probably worth pointing out too that having the same or similar header and footer on each page leads to theming which G loves as well.
I'va wooried that this may trip the dupe switch and have reduced the number of footers that are repetitive. Could be just paranoia of course ;)
Best wishes
Sid
