Tuesday, February 26, 2008

Keywords, site descriptions and meta tags - how search engines find your website and content

I'm currently updating the keywords for the pages and galleries on my website - Pastels and Pencils. When I originally set it up, the Sitekreator webware which I use only allowed one site description and one list of keywords for meta tags. At the moment Google Webmaster tools tells me that its bot keeps finding duplicate page descriptions - surprise, surprise!

I discovered recently that I can now vary the description for each page and tailor the keywords to each page. So that's what I'm now engaged in doing. However I decided to check out which were the best keywords for each page. This post (and the next one) is about that process - so if you are not an expert and want to know a bit more about keywords, meta tags and site descriptions you may well find something interesting in what follows.

My expertise: 'advanced' novice status!

I was a TOTAL novice when I started out - strictly amateur but keen to learn. I'd heard about keywords and meta tags and 'knew' they were important so I tried to learn a bit more as I set the site up. I'd really only classify myself as an advanced novice now but find that I often encounter people who know much less than me even though quite a lot of the basics is really not difficult. So what follows is my strictly non-techie understanding of and approach to keywords, meta tags and site descriptions.

I hasten to add that I only work with one form of webware, don't do the html bit on home-coded websites or advise on how to do things with other types of webware - and I only point you in the direction of where you will find tutorials or sites which are more reliable sources than me! This, as you will appreciate, is my 'get out' in case you crash your website!

Now - I don't mind experts using the comments function to correct anything I have said or to add in further comments - that's how we can all learn - with one proviso! You MUST try and use strictly non-techie language. If you must use jargon, then you MUST also explain what it means in layman's terms first.

What are keywords and meta tags?
Meta keywords
On the web, a keyword is a reference to the content and/or the type of meta element included in a given web page's HTML code to aid in the page's indexing. A keyword meta element may include several comma-separated keywords (or keyword phrases, each of which may contain several individual words)
Wikipedia: Keywords

Think of a keyword as being a bit like a signpost. Here's a very simple analogy (it's actually a lot more complicated than this). You're driving along the information superhighway and you want to find the turn-off for 'art' - well you need to be able to see a signpost for art. Then you're moving down the art highway and now want to find some landscape paintings of the Lake District in England. So you probably wait until you see 'landscape' and/or 'painting' as a signpost and turn off and then look for the signposts for 'Lake District' and 'England'. You finally end up at a website page which should contain all the relevant keywords used in your search as content. They may also be used in code as meta tags.

Unfortunately that means that all those pages which were only described in a very basic or generic way or were not labelled or described at all are probably not going to be found very easily.

There's also a twist in the tale of keywords as meta tags which not everybody is familiar with. Keywords were abused very badly by people trying to improve the profile of their web pages. Their sites would be loaded with 'heavy duty' keywords as meta tags and the search engines would be fooled and the results of searches started to become a complete nonsense.

The keywords attribute was popularized by search engines such as Infoseek and AltaVista in 1995, and its popularity quickly grew until it became one of the most commonly used meta elements[1]. By late 1997, however, search engine providers realized that information stored in meta elements, especially the keyword attribute, was often unreliable and misleading, and at worst, used to draw users into spam sites. (Unscrupulous webmasters could easily place false keywords into their meta elements in order to draw people to their site.)
Wikipedia - meta elements - they keyword attribute

People speculate about what now happens. The best guess is that meta tags probably still 'count' but that Google started to check meta tag keywords against the text content of the page. The tools they now use to check pages are becoming ever more sophisticated. My own personal 'rule of thumb' is as follows
(Experts - please note this my layman's shorthand version!):
  • If the tags, page title and page description seemed to reflect the page content then Google will be happy.
  • If the words and language are of a good standard and fit with the rest of the site then Google will be happy
  • If they don't reflect content and/or exhibit poor language structure then Google won't be happy - and might keep that page low in the index if indeed it gets indexed at all.
  • If there are very few words on a page then Google gets confused - which probably means getting a good Google ranking is difficult if not impossible.
What now follows is a brief overview of how a search engine works and how it looks at my website - and yours.

How a search engine works

Wordtracker.com provides a useful overview of how a search engine works in basic terms.
  • the spider bots 'crawl' the internet and find new pages. Spider 'bots' look for text, links and the URL. They ignore all images and formatting. They return periodically and review for changes - particularly if you have a site map loaded. (I've now got sitemaps for my website and both blogs loaded.)
  • the index software works as follows

The Index software catches everything the Spider can throw at it....The index makes sense of the mass of text, links and URLs using what is called an algorithm - a complex mathematical formula that indexes the words, the pairs of words and so on.

Essentially, an algorithm analyses the pages and links for word combinations and assigns scores that allow the search engine to judge how important the page (and URL) might be to the person that is searching.

Wordtracker.com: Keyword Basics Part 1: How Search Engines Work

  • the front end of the search engine uses query software. You or I type in the words for our search and the software then reviews all the information stored in its index (ie The search engine doesn't search the internet - it looks at what it's already got collected by the spiders and analysed by the index software. That's why sites won't appear in response to a searche until they have been crawled and indexed)
Site Description

I found out that the site description is that little bit you see underneath a site heading when it comes up in a list in Google. The description for a page is one the things that some bots look at. This is what wikipedia has to say about site descriptions - note citations are required for some of it.
Unlike the keyword attribute, the description attribute is supported by most major search engines, like Yahoo and Live Search, while Google will fall back on this tag when information about the page itself is requested (e.g. using the related: query). The description attribute provides a concise explanation of a web page's content. This allows the webpage authors to give a more meaningful description for listings than might be displayed if the search engine was unable to automatically create its own description based on the page content. The description is often, but not always, displayed on search engine results pages, so it can impact click-through rates. Industry commentators have suggested that major search engines also consider keywords located in the description attribute when ranking pages.[4] W3C doesn't specify the size of this description meta tag, but almost all search engines recommend it to be shorter than 200 characters of plain text[citation needed].
Wikipedia - meta tags - the description attribute
I use the Squidoo bookmarklet tool in my Firefox browser to add in sites to my Resources for Artists information sites. This has a very nifty arrangement where just adding in a URL address for a webpage automatically identifies both the site title and the site description for that page. Except you would not believe how many web pages out there have rubbish titles and absolutely no description - or just the first bits of text which come up on the page. I'd estimate that at least half the sites I add are described in ways which are unhelpful in the extreme.

What you can see below is what my site looks like if I attempt to add it as a link to bookmark it.

Checking your meta tags for title, site description and keywords

You can also check out how any page on the internet is described i terms of its meta tags by going to View/Page Source while viewing the page in question. Another page pops up and that gives you the visible coding for the page you are viewing. (Some sites seem to obscure this but I've found you can see quite a lot for most pages).

Go and check it with your own website right now. Now you see that bit where it says meta tags right near the top check out what it says for
  • title
  • meta name ="keywords"
  • meta name ="description"
....because that's how your site is described in search engines. (I've left out the <> )

You can see the top of the "page source" view for my own website at the top of this page - click the image to see a larger version.

Now - if your jaw is dropping at this stage just remember you have some very good company. I could name names, but I won't! ;)

Keywords / meta tags are also like 'labels' in Blogger and 'categories' in other blogs. They are the tags / signposts which help people to find a type of post. See these two posts for why labels are important when using Blogger
Which keywords?

Which keywords to use is THE BIG QUESTION. There are vast numbers of people paying other people lots and lots of money to work this out - because, as we know, keywords used as meta tags or as part of the content of your web page are the signposts are very likely to help get people to your site. So how can I (or you) compete with that?

Well you can learn quite a bit by reading around. Try reading the wikipedia page on keywords. This has a section on 'working with keywords'.
The first mistake many publishers make is to 'underdescribe' their content by using a keyword that is too general to be useful......The second mistake publishers frequently make is to not put themselves in the mind of the searcher, but to instead use keywords that are relevant to them.
Wikipedia - Keywords
When I was starting out, once I had discovered the "page source" view I started looking at well constructed artists websites - where somebody had obviously taken some care about it (I thought). I looked at how they titled their sites, how they described them and which keywords they used. I used them as a reference, they gave me an education. I also learned very fast that what looked like a super duper website on the face of it could be let down very badly by the fact that it was clear that somebody had not done the necessary work with the site description, keywords and meta tags. So that got me started.

(As an aside, what's very funny is that the Wikipedia keywords page uses itself an example and discusses what words might be best to describe that page - so I checked - and they are actually completely different!!!)

Then I discovered the fact that you can check out the popularity of different keywords - in a general sense. People often think that the number of results you get when you type a keyword into a search engine tells you how popular that keyword is. However all the search engine is doing is telling you how many pages on the Internet contain some or all of those words - it doesn't tell you how popular they are or how many people are searching on those words.

Tools exist which allow you to check the popularity and use keywords. Take a look at this page on del.icio.us - it lists all the sites which have been bookmarked and saved by people using the tag 'keywords' - and you can see how popular they are. I'm assuming popularity in this context (because del.icio.us gets used a lot by people actively wanting to look and store information) probably has some connection to how well they work. Although it might be an indication of how long they've been around - just like the early blogs captured lots of links.

Now the top one is the one used by Yahoo and is very popular and recommended by my webware - but it periodically seems have a habit of conking out as it gets overloaded due to usage.

So take a look at the next best thing - the external version of the Google Adwords tool. Google Adwords are really just keywords used for advertising and the placement of Google Ads. However, you don't need to be subscribed to Google Adwords to use the tool.

You can also use Google Trends to check out what people are searching on - but this is a tool which is much more relevant to people whose sites are very newsy and up to the minute!

The problem for artists

As you can imagine lots of images with esoteric / punter friendly titles and very little if any other text on a page might have a problem presenting enough data for the search engines spiders and index software.

So an artist's website is always going to create a problem in terms of being easily found because of the nature of the content of most pages. That's fine if you are happy that the only people who will find it are those that you personally give the website address. It's not so good if you'd like other people to find it too.

The same issue arises for those auctions and blogs which aim to to sell artwork.

How to remedy this really needs to be the subject of another post.

At the moment, I think it's
probably worth my while going through my website, reviewing my page titles, my content for keyword use and then changing all the page descriptions and keywords for each page of my website to get a better match with content.

The next post

In the next post - later this week I'm going to highlight how to find and choose keywords and how to change them on your website using the tools available - within the provisos set out above.



  1. "As you can imagine lots of images with esoteric / punter friendly titles and very little if any other text on a page might have a problem presenting enough data for the search engines spiders and index software."

    This is precisely what the alt attribute on the img element is for. This attribute allows you to describe the image and is responsible to the text you'll see in modern browsers when an image fails to load (provided the site author has provided for it). It is also what screen reader software speaks aloud to users of such technologies.

    An example:
    <img src="web_Pink-petals#2.jpg" alt="A colored pencil drawing entitled Pink Petals. Colored Pencil on Arches Hot Press, 6 x 6" © Katherine Tyrrell" />

    I've left out the additional attributes on the image in the example (width, height, etc) for brevity. Obviously, the more descriptive you get with the alternate content the better.

  2. Thanks - I told you I was only an 'advanced novice'!

    Although I knew that including information like that was a good idea it's not actually what's shown up in the alt attribute at present - so I need to find out how to do that in blogger.

    So I guess I need to go and view this video by Matt Cutts, the head of Google's webspam team? It's all about Using al attributes smartly

  3. I just tried the keyword tool and used the website content option. It returned a long list from which I could select the keywords I wanted to use.

    After selecting words applicable to my website, Google put them in a list I could download as a text file, free of charge. I now have a handy-dandy list to use to update my meta tags and keywords.

    Thanks Katherine!

  4. Well I've tried it and Blogger doesn't seem to like it at all.

    Anybody got a more detailed reference which works with Blogger coding?

  5. It looks like you'll have to modify the markup (by clicking on the "Edit HTML" tab) and add the value of the alt attribute yourself. Handily, the attribute itself is already there, its just lacks a value.

    View the source of my test blog to see that it does take, you just have to do it by hand.

    I'd recommend petitioning Google/Blogger to add support for the addition of alt and title attributes in their image uploader, especially if you're uncomfortable with editing markup.

  6. Thanks. I edit the html all the time - but very rarely alter the the image html. I was trying to alter the image for the one posted yesterday, which you used in your example, but it didn't work. I'll have to see if I can work out why not.

    I agree - an alt image caption as part of the upload image screen would be very helpful

  7. I'm an advanced novice, if such a thing exists. ;) These things change and evolve so much that you're only "in the know" for a short while unless you keep up with the techie forums and things.

    Can I recommend, highly recommend, a FREE search engine optimization course via a Yahoo group. It's intense and you will have to do work to identify you niches, right keywords and all, but it's run by an expert and well worth it. Very general business oriented so artists will have to think a bit laterally. (She also has an e-book of it all if you're too impatient for the online class!)

  8. Thanks Tina - it's always good to get links to courses that artists have done and can recommend as being relevant for and accessible by artists.

    Every time I think I'm getting on top of something they go and change it! ;)

  9. a very helpful post Katherine for someone who has a website in desperate need of pruning and tidying!


  10. Oh dear. Jaw dropping! I am definitely one of your encounters with 'someone who knows much less than you do about meta tags'.
    This is a fantastic post, Katherine. When my head stops spinning I will attempt to pull my cyber socks up.
    Yours ever gratefully...

  11. I'm sure my blog fails miserably in many ways as I'm hopeless at these things but I do think if a blogger is going to a lot of trouble to get more visitors then they should at least spend more time on their site descriptions, so many are woeful. I have trouble putting pyseudonyms to blogs and I like to know at least which continent they live on. To them, they may be irrelevant or obvious but it should be an opportunity to give an overview of what the blog is about, like a welcome as you walk in the door not left like a stranger wondering what on earth is going on!

  12. Great article, Katherine - a really good start here. Probably the most important point I can add to your article and S.G. Chipman's comments is that the things search engines (and especially Google) place the highest importance on is actually the content visible in your page.

    HTML allows for clear, meaningful tagging of headings in the page using the H1, H2, etc. tags - search engines will look to the content in these tags (esp. H1 and then descending in order of importance through the numbers) when it's ranking your page. Headings being properly marked up (I know this isn't always controllable when using a content management system), being meaningful and the first paragraph of text on the page likewise are the things that will have the most affect on your page ranking. When reviewing a page, keep your keywords for that page handy and see if you can work your headings to include one or more keywords in there. Don't overdo it though! Remember that the search engine is also checking for meaning and your users also want something that's readable. But it's important to avoid using "cute" headings. If it doesn't fully convey the sense of the page or the section that follows, it needs reworking. Similarly with the first paragraph - keep it clear and descriptive of what content people are going to find on the page. Think about it this way: if you came to your page, not knowing anything about it, and all you were allowed to read was the first heading and the first paragraph, would you know what the page was about and would you want to read more? If the answer is "no", then some revising is needed :-)

    Link text is also important both for search engines and to ensure your site is accessible to users who may be accessing your site using screen readers or other non-mainstream technologies (these two always go hand-in-hand because search engines can't "see" your page, only your code). Don't use generic links (e.g. "click here" or "read more") and instead use text which helps describe where the link is going. This helps because people using screen readers have an option to "scan" all the links in the page, which just reads out the text in the links (imagine: "click here", "click here", "click here"), and also because search engines do check this text when ranking the pages the links go to. It's nowhere near as important as headings and the first paragraph, but it does matter.

    I could rabbit on about this all day (I wrote a paper on this for a large professional services firm a little while back) but shan't bore you any further. I hope this helps. And I should say, for others reading this post (well done for getting to the end, too - bravo!) that Katherine's headings and links are excellent examples of the right way to go about things!

  13. Thanks for that very useful contribution Caitlin.

    I'm glad you put that extra comment at the end - I was just about to say I've already learned all that and try to implement it every day on my blog - which has made overhauling a website that much easier!


COMMENTS HAVE BEEN CLOSED AGAIN because of too much spam.
My blog posts are always posted to my Making A Mark Facebook Page and you can comment there if you wish.

Note: only a member of this blog may post a comment.