July 15, 2004 | View Comments (21) | Category: Web Mastering
Summary: How important and effective are meta-tags nowadays?
Meta-tags are easily one of the most overlooked aspect of any of my sites. I finish them in about 10 seconds and then I move on to the fun stuff. At the beginning of the web the use of meta-tags was seen as integral for your strategy of achieving high search engine rankings. If you studied the code you would see description tags that were over 500 words filled with junk. It got so bad that search engines decided to ignore them and actually focus on the content of the site (good move).
However, meta-tags are still useful and you never know when they might become the “in” thing all over again.
On Whitespace I use a total of two meta-tags on the homepage.
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
<meta name="robots" content="noindex, follow" />
Those seem to work fairly well for me and to be honest I don't know how hurtful it is for me not to include more tags. Case in point if you header over to Zeldman's site you will see that he uses a total of 9! tags and it makes me wonder how useful they are.
Really I am no expert in the meta-tag field so if you know some things about them I would love to hear your thoughts. Is it useful to still include a lot of them like Zeldman does? For example, the author, description and keyword tags or you better off just sticking with the ones that the search engines focus on?
Trackback URL: http://9rules.com/cgi-bin/mt/mt-tb.cgi/281
I think the GeoTags for describing your geographic location are an interesting and potentially useful use of metadata. I tend to be pretty sparse with my use of meta information in general, though. It seems the usefulness of any meta tag that can be abused (e.g. description, keywords) approaches zero over time.
Tom Werner (http://www.mojombo.com)
To me there I only a handful of meta tags I would describe as being useful. Content-Type is a given, becuase I like validation. Robots is great for telling search engines what and what not to cache.
Other than that I see a lot of them as being uncalled for. In a older web, meta-tags relayed information which was supposed to be useful. However, as you stated useful, soon became untruthful.
Copyright and Author are others that might be useful, however I do not know there exact purpose. I assume that at one time it was thought to be a form of digital signature that never got implemented into mainstream browsing software (I;m not sure about this). The only real use I see to them is to prevent design theft, when a less knowledgable person feels the need to jack your design and lacks the understanding of actual HTML code.
Maybe that hits a little to close to home for me, and I am being spiteful. And I appologize for that, however, as far as those tags, and the various other meta tags you see floating around. I'm not sure if I would call them useful, or just call them existing.
It's funny you point out the meta tags on Zeldman's site. And I hope to hell that every single meta tag included in there is somehow useful, or else he just contradicted the first two chapters of designing with web standards. I hope it isn't creating a downward spiral of cost effectiveness, in a redudnat form of deeply nested meta tags :-O
Ryan Latham (http://www.worldoneweb.com)
(sorry for the lengthy reply...)
What we need to do is distinguish document, site and technical meta tags. Not all of them are useful.
Document meta tags is very important and it allows you to do more with your content.
Most importantly meta-data is not about adding a link or meta tag, which is in most cases irrelevant. But take this entry as an example. You got two very important types of meta data: The date and what category it belongs to.
The date allows you to categorize entries by date, or more precisely it makes it possible for your back-end system to do so. It also allows your reader to know the freshness of the entry. Another way to increase usability of your site is to show related entries - and a date is an important way to ensure that you do not show out-of-date entries.
The category allows you make classification, to divide you site into sections. It also enables your readers to know what the entry is about. This entry is about "Web Mastering" (although it might have been better to use Information Architecture).
Another thing is keywords. Many people talk about adding keywords, but few actually use them correctly - so they are left in the waste-of-time pile. But keywords are immensely important. What it allows you to do is to ensure that the article is exactly about what you are looking for.
To give an example: On this site you have a search function. If the search function used document keywords to rate the results you will only get relevant results - whereas if you search the whole content body you will often get irrelevant results.
Consider two entries. One is this one, another one is about "Logos and Visual Identity". If you do a search for "meta" both of them will be found, but only one of them would be relevant.
- Author, tells you who wrote it
- Headline, is the shortest description
- Contributors, ... Well guess :)
- Coverage, is the scope of the entry
- Description, or summary is a more in-depth information of the entry (excellent for showing in search results)
- Format, can be used if entries are published in more than one format (html, xml, word, pdf, rtf etc.)
- Language, allows you to... well identify the language
Etc.
Meta data is about helping find what they are looking for, tell them what is about, when it was about it and who made it.
In terms of meta-data (x)html tags this is only important when we are talking interoperability - i.e. sharing content between systems. Search-engines are one way, but since they only use description there is really no point adding the rest. I have in past made a few systems where the (x)html pages did include all the meta-data, simply because one server was used as a publishing server, then we had 4 external servers picking up the relevant pages (this was later changed to XML).
Thomas Baekdal (http://www.baekdal.com)
Also note that Zeldman first started his site in 1995; maybe old habits die hard.
I use two: one defining the charset, and one defining the generator as WordPress, for statistical reasons.
Benjamin (http://www.bluism.com)
Why is Content-Type given? If you set it serverside, the browsers actually do not need it. Jusk ask the USER_AGENT-string if it is a validator and the write the Content-Type meta-tag.
And regarding the meta tag robots, why not let the search engines index all of your pages? Or simply just use the robots.txt and make some pages allowed and some disallowed.
Martin Hintzmann Andersen (http://www.hintzmann.dk)
Meta tags *are* important.
If you're prepared to 'future-proof' your documents by creating interoperable, valid XHTML documents, then why aren't you also creating the proper meta tags?
When we talk about 'semantic' HTML we are talking about using the inherent properties of the elements to enhance the document text, and better define its meaning - so that both humans and machines can understand it. (h1, h2, for heading text)
They are there to provide additional information about the content of the document, (and without wanting to come over all purist), surely the point is to produce *perfect* (X)HTML and assume that search engines will catch up?
Build it right first time, otherwise, you have to go back and rebuild it all later...
In the meantime, I defy you to produce a well authored document that is negatively weighted by search engines.
pid (http://www.pidster.com)
If meta tags are no longer relevent, why haven't they been deprecated in one of the 3 intervening versions of the language?
From the HTML 2.0 specification documents.
http://www.w3.org/MarkUp/html-spec/html-spec_5.html
--snip--
The META element is an extensible container for use in identifying specialized document meta-information. Meta-information has two main functions:
* to provide a means to discover that the data set exists and how it might be obtained or accessed; and
* to document the content, quality, and features of a data set, indicating its fitness for use.
Each META element specifies a name/value pair. If multiple META elements are provided with the same name, their combined contents--concatenated as a comma-separated list--is the value associated with that name. (13)
HTTP servers may read the content of the document HEAD to generate header fields corresponding to any elements defining a value for the attribute HTTP-EQUIV. (14)
--snip--
pid (http://www.pidster.com)
I assume the reason he uses them is because 1 or more search engines still use "meta keywords" and "meta description" and the others, thus making his site accessible to those that use those search engines.
He's just ensuring that he doesn't "lose" some of his audience I guess. I use google all the time, but i'm sure there's lots of people who prefer some other search engine like "lycos" or "dogpile".
I use the same meta tags that Zeldman does.
Robert Lofthouse (http://www.ghxdesign.com)
I try to always use meta tags to include my name somewhere, so that in theory, even if my name isn't mentioned anywhere else on the site. I can search for my name and get results including all the pages I've ever worked on.
I thought it was recommended that you declare content type on the server so I avoid that one.
I agree that even if some search engines don't use them, some probably do and it certainly doesn't hurt to have them. Providing more information about your site is usually a good thing right?
Sunshine Lewis (http://sparklit.sparkalyn.com)
Wow, great information so far everyone. Thanks a lot. For the record I use the robot meta tag so that the index page does not come up when people do searches as it is constantly changing.
Scrivs (http://businesslogs.com)
I had a bad run in with the Content-Type meta-tag. It just didn't work when displaying UTF-8 encoded webpages that were localized in 4 languages (excluding English). I was told that it had to be the very first thing the browser sees, and placed it as the first element in the tag, but no go. To this day I don't know why this is so.
The problem is fixed if you send the Content-Type in a HTTP header, and that is actually the right way to go about this. The meta tag is there when you really cannot send the appropriate header with server-side scripting (in PHP, just use something along the lines of header("Content-Type: text/html; charset= UTF-8"); ).
Cheah Chu Yeow (http://blog.codefront.net/)
These are some of the meta tags I've used: dc.creator.name, dc.creator.e-mail, dc.publisher, dc.subject, dc.language, tgn.id, tgn.name, icbm, geo.position, geo.placename, geo.country, dmoz.id
Besides these (possibly) being used by different services, like GeoTags, Syndic8, etc, I always think I might find a use for them in my own applications.
As for setting the Content-Type in the document using a meta tag, in addition to doing it on the server, see what Joel has to say: http://www.joelonsoftware.com/articles/Unicode.html
Pete Prodoehl (http://rasterweb.net/raster/)
I won't list them all here, but we use loads of meta tags on one of our sites - perhaps it is because many of our clients are Government of Canada departments -- the GoC has mandated metadata, and one of our goals on our site (WATS.ca) was to demonstrate that we understand the importance of meta data in general.
I see it generally being less important for human consumption, but more important for machine consumption. If something in the metadata is important enough for humans to know about, it is up to either the browsers to display it (not likely in this generation of browsers), or for the author to render it in the page.
When we built the CMS behind the site we added a metadata repository for each resource. That data is added to the documents, but I see more and more everyday that it is mostly used by me when I am coding new functionality (like the Related: section we have on each of our articles), and possibly for other future scripts/programs/whatever - whether we are talking about adding functionality to the site, or about search engines or other "bots"
feather (http://boxofchocolates.ca)
I haven't seen anyone mention meta ratings. This can be an important tag if your site contains content not suitable for minors, or contains content which a netnanny type system may consider restricted but shouldn't be.
There's the meta ratings tag, which works rather like the MPAA ratings.
Then there's the more sophisticated system, PICS. To see what a label would look like, here's a generator.
Also P3P is important, since newer browsers use that to determine the level of permission your site has in setting cookies and the like. It's a pain in the arse, but sometimes it's the only way to get stuff working properly. More info here.
Of course, PICS and P3P both can also be sent (and probably work much better as) a part of the HTTP header.
Another one I don't see is META NAME="GOOGLEBOT" CONTENT="NOARCHIVE", which could be useful in some situations, regardless of how much it annoys me (I often view the archive copy of pages I find in google because they tend to load much faster)
JC (http://thelionsweb.com/weblog)
META Keywords and Description are an ongoing debate at my day job. A few lower tier search engines are META only. ExactSeek, for instance, uses ONLY the META tag data for listing. We tend to use the META keywords list to run WebPosition reports to evaluate ranking of pages, so the process of working with the client to draft the META serves these two functions: placating the META only search engines and composing the list for running WebPosition. In addition to the typical META, we include the Pragma tags to prevent browser and server caching: some of our clients seem to ALWAYS see cached content no matter what we do, so we use the Pragma to help nip that problem in the bud.
Wayne Hastings (http://www.hastingsinteractive.com/)
META tags are about providing machine-readable meta content about your pages. The target agents who will use these could be UAs like browsers or some addon appliction like the P3P bird (this looks up the P3P statement and highlights whether the privacy policy of the site matches those preferred by you) as well as search engine bots or some other software.
The idea of using these is to provide correct data in correct format to automated information retrieval and classification systems (Search Engine bots are just one example).
One should never use these thinking some human will read them.
One should not use them because of what I call 'search engine bot relevancy' -- the whole idea is unethical and frankly, stupid. When I come across SEO-experts who discuss how to exploit meta information to get higher search engine rankings, i am amused by their desire to do so simply because it is doable. Hence, irrelevant search engine results and what not. Maybe one day search engines will simply stop reading META data.
But the point is that the meta data is not ONLY for search engines. Thinking about meta data in terms of search engines is a flawed approach. There might be an application out there that is merely collecting the meta data and does not even parse the <body>. You'll end up classified wrongly. If your meta content is missing, you'll not be part of the database at all.
As a web-page engineer and owner, a person should simply ask a question whether they want to make their page's meta information available to machines that collect/use this information for some usable task. You should remain honest and to-the-point in providing this information.
Humans--they might reach your site because of your meta content juggling, but will stay only if your real human-readable content is what they are looking for.
Use os META tags for instructions to agents (Pragma, et al.) is a completely different thing -- these do not provide any information so I think they're out-of-scope in reference to the main post.
Abhay S. Kushwaha (http://blog.kushwaha.com)
I think two of the important tags Zeldman (and others) implement is the "SmartTags" and "imagetoolbar".
I don't know if smarttags is really still an issue anymore, but I know the image tool bar - a nice little menu which pops up on top of images - still is.
Just a couple of methods to keep IE6 under control.
Mark (http://www.lightpierce.com/ltshdw)
The current mode of thinking is to match your keyword phrases and description to your content, but try to focus on only most relevant keyword phrases to that page. Don't use any single keyword more than three times.
This site has good forums on SEO topics:
http://www.jimworld.com/apps/webmaster.forums/action::topiclist/forum::seo-101/
In general you should still use Keywords and Descriptions meta tags at the very least. Most search engines look at description meta tags, but they may rank it differently according to their algorithm. As of right now, Google doesn't read the keyword meta tags anymore, but many smaller search engines still do.
Scott (http://www.scottmaira.com)
Here's a meta-tag I wish existed... the Non-Keywords metatag, to help exclude particular keyphrases from showing up in search results... These aren't the pages you are looking for
Anyone know of anything like this? or come up with other ways to accomplish the same?
feather (http://boxofchocolates.ca)
I feel it's a positive process when I analyse the pages I create and add keywords and description meta tags. It helps me stay on track, and provides a framework for a range of different types of key word/phrase analysis. These meta tags also help to remind me what I was targeting when I return to a page a while after I originally published it.
Matt Williams (http://www.ecru.co.uk)
Keep track of comments to all entries with the Comments Feed
©2003-06 9rules, Inc. | Thoughts Feed | Notes Feed
Amazon Stores: Powertools | Jewelry | Books | Health | Landscaping | Kitchen | Movies
#1
Last checked, Google (and Yahoo) uses meta-description for search results, provided text from your query can be found in the meta-tag.
Fwiw, it would seem that the content of the meta-desc should also be found in the body of the page* - sticking extra stuff in the meta-desc that is not contained in the body does not work (in cases where I've tried).
* not word for word, but the important words should be found in the body section.
Mike P. (http://www.fiftyfoureleven.com/sandbox/weblog/)