Revisiting The Meta Description Tag
A major tour of how page descriptions are currently formed at several major search engines. The meta description tag largely remains useful.
A major tour of how page descriptions are currently formed at several major search engines. The meta description tag largely remains useful.
How do search engines form the descriptions for your web pages? In the wake of the recent Google redesign and Yahoo’s use of its own technology, I thought a revisit was in order. I’ll cover several different search engines, and you can jump to each one directly via the links below:
Jump To:
Google – Yahoo – AltaVista
MSN – Ask Jeeves/Teoma
In general, Google will try to make descriptions (what it calls snippets) by automatically extracting text from the web pages it lists. The text it extracts from the page will usually include the actual search term used that brought up the page.
Huh? Take an example of a search for shoes, where the first site listed has this title and description:
Shoes On The Net – Men’s Shoes, Women’s Shoes and Children’s Shoes …
Shoes On The Net – Shoes, Fashion Updates, A Personal Shopper and The Worlds
Largest Shoe Mall. … As seen in Allure magazine. Kangaroo Shoes – For Men and
Women. …
www.shoesonthenet.com/
Where’s the description coming from? Google looks on the page and finds the first reference of the word “shoes”, then pull back the text around this. In this case, that reference comes from text in the meta description tag, which looks like this:
Shoes On The Net – Shoes, Fashion Updates, A Personal Shopper and The Worlds
Largest Shoe Mall.
Of course, the Google description is longer than what’s in the meta description tag. Notice the ellipse in the Google description, the … part that follows the text that comes from the meta description tag? That indicates that the next extracted text is pulled from elsewhere on the page. In this case, Google has selected additional text surrounding another area where the word “shoes” is used on the page.
By the way, if you aren’t familiar with the meta description tag, then see the How To Use HTML Meta Tags page, which explains it in depth.
Just using the meta description tag doesn’t guarantee the text will always be used by Google. For example, look at the description for this site listed first at Google in a search for showtime dance shoe:
Dance Shoes: Showtime Dance Shoes and Dancewear
… Showtime Dance Shoes was founded in Atlanta, Georgia over 30 years ago. … Flash
Catalog. ) 2000-2003 Showtime Dance Shoes all rights reserved.
www.showtimedanceshoes.com/
In this case, the word “shoe” doesn’t appear in the meta description tag. Because of this, Google bypasses that text and instead looks for the first reference to “shoe” elsewhere. This is found down in the body of the page.
In contrast, look at how this same, exact page is listed when found through a search on shoes:
Dance Shoes: Showtime Dance Shoes and Dancewear
Ballroom dance shoes and Chrisanne Dancewear for the social or professional dancer.
We carry Supadance, International and Freed of London dance shoes. …
www.showtimedanceshoes.com/
In this case, the meta description tag does get used, because the word “shoes” appears within it. To be clear, because the meta description tag had the first reference of the word “shoes” that Google found on the page, iit had a high likelihood of being used.
Should you now go out and try to put every conceivable keyword into your meta description tag, in an effort to “force” Google to use it? No. It will likely make your description look worse. For more advice on this, see this good, recent article by Jill Whalen on why you shouldn’t go too far.
Google has long supported the meta description tag as a solution for making descriptions in cases where a page lacks body text, and this continues today. Look at these descriptions for top ranked pages from a search on shoes:
Welcome to Vegetarian Shoes, supplying animal friendly footwear …
Vegetarian Shoes supply animal friendly footwear and clothing for vegetarians and
vegans worldwide.
www.vegetarian-shoes.co.uk/
:: GLOBE SHOES :: skate and surf footwear and clothing
Manufacturer and distributor of Globe skate shoes, Trans Element casual shoes and
streetwear apparel. Includes product information …
www.globeshoes.com/
In the first example, the page is a frame that lacks any real body text. However, since it has a meta description tag, that’s what Google decides to use.
In the second example, the page is all Flash content, providing no body text for Google to form a description from. But since a meta description tag is employed, that’s what Google uses.
What happens if a page lacks both body copy and a meta description tag? Then Google might rely on a description of the page if it is listed in the Open Directory Project. Look at this example, again from a top ranked page for a search on shoes:
DCSHOECOUSA
Team profiles, a company history and pictures of all the crazy shoes they’ve put out.
www.dcshoecousa.com/
That page appears to only contain Flash and JavaScript content, giving Google no body copy to lock onto. It also has no meta description tag. So where’s that nicely written description coming from? My guess was the Open Directory, which proved correct. Here’s how to confirm that, if you’re in the same situation.
Head to the Google Directory, Google’s version of the Open Directory. Search for the URL or page title of the page you saw listed in Google’s web search results. In this case, I searched for www.dcshoecousa.com. The page came up in the directory results. You can see how the Open Directory’s human-written description appears right in this directory search result.
You can also clickthrough on the category link shown, if you want to see how the page appears within the category. Do it in this case, and you’ll discover why it might be better to search for the page’s URL rather than page title.
When looking at the actual category, you’ll see that the page title is now DC Shoes (as the human editor would have written it) instead of DCSHOECOUSA, the title Google made by reading the page’s title tag. Had you searched by title, you might not have found the ODP listing.
By the way, until Google’s recent redesign, any site that was listed in the Open Directory would be flagged with both a category link and an Open Directory description in addition to the Google automatically created snippet, when it appeared in Google’s web search results.
How about an example? Say you searched for origami before the change. One of the top ranked pages would have looked like this:
paperfolding.com
A fun site to learn about origami, the art of paperfolding! Learn to
fold, see the origami insects and dinosaurs, learn about origami …
Description: Diagrams, history, relationship with math, gallery, and
book reviews with cover shots.
Category: Arts > Crafts > Origami
See the two last lines, the Description and Category ones? These were coming out of Google Directory. Google was showing you matches from across the web, then looking to see if those matches were also shown in its human-powered directory. If so, then it supplemented its automatically-generated description for the page with the page’s description from the Google Directory and a link to its home category.
Now the same site just looks like this:
paperfolding.com
A fun site to learn about origami, the art of paperfolding! Learn to
fold, see the origami insects and dinosaurs, learn about origami …
The supplemental Google Directory information is no longer shown.
Here are a few last things to keep in mind about Google and descriptions.
First, Google won’t show descriptions for what it calls partially-indexed pages. These URLs haven’t been spidered by Google, which is why you see no descriptions.
I prefer the term “link-only listings.” That’s because Google only knows about these URLs from what other people say when linking to them, rather than from having read the pages themselves.
To see an example of how these look, see my past article, The US White House & Blocking Search Engines, the “What’s Really Blocked” section.
Next, be aware that you can tell Google that you don’t want it to make descriptions for your pages by employing its unique nosnippets meta tag. You can learn more about that from Google here. Be aware that if you do this, you’ll end up with NO description at all for your pages.
Over at Yahoo, one of three major elements may be used for your description, depending on what Yahoo’s search system determines is best for the user. These are Yahoo Directory descriptions, the meta description tag and page copy extracts. Let’s take each one, in turn.
Like Google, Yahoo also has a human-compiled directory of web sites. This is the Yahoo Directory.
When you search at Yahoo, it will check to see if any of the pages listed from its web crawler also live within its Yahoo Directory. If so, then it will use both the human-written title and description of the page from the directory to describe it in the web search results.
This is something Yahoo has done since its October 2002 redesign. The only recent change since then is that the little red arrow that used to appear on the category line disappeared toward the end of last year, if I recall correctly. And, of course, Yahoo is using its own crawler-based results now, rather than depending on Google.
Here’s an example of a directory description being used for a top-ranked site for shoes:
Zappos.com
offers brand name men’s and women’s shoes from sizes 3-18.
Category: Footwear Retailers
www.zappos.com/
Zappos.com does have a meta description tag, but Yahoo has ignored it. Instead, the description shown is how Yahoo’s editors have listed the page within the Yahoo Directory category where Zappos.com is also listed. That category is shown below the Zappos listing. If you were to click on it, you’d find the source of the description you are seeing.
In general, if a page is listed in the Yahoo Directory, I’ve found that Yahoo almost always seems to favor using the directory description rather than a meta tag or an extract.
What if you aren’t in the Yahoo Directory? Then Yahoo might use your meta description tag, as seen in this case in a top result for a search on shoes:
Underground England – Boots Shoes Creeper Sneakers Retro Gothic
Boots, Shoes, Creeper, Sneaker, Winklepicker, Baseball Boots – THE alternative British
urban lifestyle brand inspired by the British musiculture … Beat boot, Boots, Belt, Biker
shoes & boots, Bowling shoes, Boogie Creepers, Boxing boot, Brogues … Creepers,
Buckle, British footwear, Blue Suede Creeper shoes, boys footwear, Belts. …
www.piccadilly-shoes.co.uk/
See this portion of the description?
Boots, Shoes, Creeper, Sneaker, Winklepicker, Baseball Boots – THE alternative British
urban lifestyle brand inspired by the British musiculture …
That comes directly from the page’s meta description tag. Then see the ellipse portion, the … part? As with Google, this indicates that the remaining portion of the description was extracted from body copy on the page.
In some cases, you may see a page that looks like it has a description coming out of a meta description tag. However, if you look at the page, you might not find the tag there.
What may be happening is that you are looking at content coming in through Yahoo’s Site Match Xchange trusted feed program or through its non-profit feeds. In that case, the description is obtained as part of that feed’s meta description information, rather than from the content of the page itself.
For example, look at this top result for a search on shoes:
Shoes – Womens Shoes, Mens Shoes, Childrens Shoes – Shoes.com
Womens Shoes, Mens Shoes, Childrens Shoes, shoes, sandals, boots, clogs, slippers
and accessories at Shoes.com. Free shipping & returns, no tax. Shoes.com offers a
great variety of Adidas, Asics, …
www.shoes.com/
Where’s that description coming from? Not from the actual page. Instead, the description content is being found by Yahoo’s review of the trusted feed information. That will likely be very similar to what’s on the page, but it won’t be exactly the same.
Have no meta description tag and not listed in the Yahoo Directory? Then it’s very likely that your description will be formed by pulling text automatically from your body copy. Here’s an example from a search for bridal shoes:
Bridal Shoes
… Choose from a stunning selection of bridal shoes from top designers, from Vera
Wang … Stuart Weitzman Quadro Bridal Shoes. Available for $205.00 per pair at
Bellissima Bridal Shoes …
www.weddinggazette.com/category/006300.shtml
By the way, at one time back in the days of Inktomi, I recall paid inclusion guaranteed getting your meta description tag used. Yahoo says that’s no longer the case, under its new program. Even if you are in one of its paid inclusion programs, it will automatically determine what is the best description to form for the user, rather than go with meta description information.
I’d actually prefer they stuck with the old system. It seems like one of the benefits of doing paid inclusion is that you should at least have more control over your page description. As long as you aren’t misleading, if you pay, you should be able to say what you like.
Yahoo-owned AltaVista now uses Yahoo-powered results. However, AltaVista does NOT use Yahoo Directory descriptions, in the way that Yahoo does. Other than that, things appear to work exactly as with Yahoo except for one last exception.
What happens if you have no body copy and no description? Then oddly, Yahoo makes use of the rival Open Directory’s description. For instance, look at this top-ranked site for shoes at AltaVista:
DCSHOECOUSA
Team profiles, a company history and pictures of all the crazy shoes they’ve put out.
www.dcshoecousa.com
Compare this to the same description at Yahoo for shoes:
DC Shoe Co USA
makers of professional quality, technical skateboarding shoes and apparel, in addition
to snowboarding boots.
Category: Athletic Shoes > Brand Names
www.dcshoecousa.com/
In the first case from AltaVista, the title comes from the page itself and the description comes out of the human-powered Open Directory, where the site is listed within the Skateboarding Clothing category.
In the second example from Yahoo, the title and description come from the page’s listing in the Yahoo Directory.
Another oddity I noted was that the listing for Nike.com at Yahoo for shoes was in English while at AltaVista for shoes, it was in German.
What about Yahoo-owned AllTheWeb? I looked over there only briefly, as the service isn’t used much. The situation seems to match that at AltaVista.
MSN Search is powered by Yahoo, but as with AltaVista, MSN does NOT make use of Yahoo Directory descriptions.
If you use a meta description tag that contains the word someone searched on, then MSN descriptions do seem to begin with content from that tag, as is the case with Yahoo.
No tag — or does your tag not have the search word in it? Then your description seems formed from by extracting body copy text. However, Yahoo’s extract may not match those at MSN. For example, compare this top listed site for a search on clipless pedals at Yahoo to MSN:
clipless pedals
Cycling Forums > Tech Corner > Cycling Equipment > clipless pedals. Register
with CyclingForums.com. If this is your first time here then please take the time
to register. … out on the upward pedal stroke. Clipless pedals are ESSENTIAL
for mountain biking, especially technical riding … almost like skiing w/ my
clipless pedals: cruise down, put weight on …
www.roadcycling.net/t66937.html
clipless pedals
Tour of the Southern Grampians – Start List View New Forum Topics Today’s
Forum Topics New Usenet Topics Today’s Usenet Topics Register | User CP |
Members | Calendar | FAQ’s | Search | Reviews | …
www.roadcycling.net/t66937.html
OK, the first example comes from Yahoo. Notice how the extract uses body copy text around various occurrences of the words “clipless pedals.” That’s what we’ve come to expect in terms of extracts.
At MSN, this doesn’t happen. Instead, in this case, the description is being formed by the first visible text on the page, regardless of whether the terms occur in that text. That’s an old-fashioned way of making descriptions used in the earlier days of search engines.
(FYI, the “Tour of the Southern Grampians” part of text no longer appears on the page, as it has since changed since being spidered. But the Yahoo cached version showed that it was there before).
Using the first visible text for descriptions isn’t always the case at MSN. For example, most of the descriptions I reviewed for clipless pedals seemed to be of the more intelligent “extract around search term” variety. Yet, even these type of descriptions still didn’t always match what was at Yahoo. In short, don’t expect what you see at Yahoo to be exactly what you get at MSN.
What happens if there’s no body copy and no meta description tag? The page simply won’t have a description, as this example for a search on shoes shows:
DCSHOECOUSA
www.dcshoecousa.com
Finally, be aware that all bets are off if MSN does what’s becoming an infamous reversal to using LookSmart data.
MSN said last month it would definitely no longer use LookSmart listings in its keyword-driven search results last month. This was after earlier reports of LookSmart reappearing within MSN Search.
Nevertheless, the old-style LookSmart directory results — ones coming below a “Web Directory Sites” heading — were definitely present when I did a search earlier this week for “shoes,” while working on this story.
If you see your site listed under this heading, then your description is coming from how you are (or once were) listed with LookSmart.
Results at Ask Jeeves come from Ask Jeeves-owned Teoma. Similar to Yahoo, descriptions may come from extracting body copy, using a meta description tag or using a human-written directory description, depending on what’s seen as best for the user, according to Ask Jeeves.
Below is an example of how the meta description tag may be used, as seen in this example for shoes at Ask Jeeves:
Designer Dress Shoes by Prada Gucci, Camper Tods, Mens and Womens
Italian designer dress and sport shoes, including prada, gucci, tod’s, camper and hogan.
Women’s and men’s styles are available…
From:www.shoesavenue.com/
That comes directly from the page’s meta description tag. As with Google and Yahoo, having the search term in your meta description tag is likely to help the copy be used.
Did you fail to provide a meta description tag? Then Ask Jeeves may produce a description from your body copy, looking for the first reference of your text. For instance, here’s how Timberland appears in a search for shoes:
Timberland
My Account | Customer Service | Stores | Search … 0 items. Men’s Footwear. Casual.
Office. Boat Shoes. Sandals. Boots. Multi-Sport. Hiking.
From:www.timberland.com/
In that example, Ask Jeeves looked for what appears to be the first body copy reference of the search term “shoes” and grabbed the text surrounding it to make the description.
What happens if a page has little body copy, or perhaps the body copy doesn’t mention the search word? Then Ask Jeeves may rely on Open Directory listings. Look at how Payless appears in a search for shoes:
Payless ShoeSource – Womenbs Shoes, Menbs Shoe…
Shoes for the entire family.
From:www.payless.com/
That description, “Shoes for the entire family,” is simply how Payless is described in the Footwear Shopping category at the Open Directory. That’s also why Shoes.com appears this way in a search for shoes at Teoma:
Alan’s Shoes
Brand name shoes for men and women.
www.shoes.com/
I’m guessing that perhaps Shoes.com was once known as Alan’s Shoes and the Open Directory hasn’t updated its listing within the Footwear Category. Regardless, it shows why it makes sense to continue paying attention to your Open Directory listing. It may impact how your page description appears.
Also note that in this example, Teoma made use of both the title and the description from the Open Directory, rather than the description alone.
To sum up, if you want to control your page descriptions, these are the steps to follow:
Use a meta description tag, and ensure that the most important terms you hope the page will be found for appear in the tag (focus on two or three phrases).
Review descriptions of pages listed with the Open Directory or the Yahoo Directory, and request any correction of major factual errors. Don’t expect a rewrite to make your description more “attractive,” however. Editors aren’t likely to do this.
Don’t pull your hair out. With the search engines using so many different description-generating options, you simply don’t have the control you might like over your page descriptions, unfortunately.