Creating Quality Metadata for Your Web Site
Metadata on a Web site includes such items as classification (how content is organized by subject matter), date of publication, keywords, headings, summary, author name, and copyright. You need to identify metadata on your Web site to ensure that:
- every document is properly classified so it can be found quickly
- all relevant legal and administrative information on a particular document has been collected
- you maximize the chances of a document being indexed appropriately by internal and external search engines.
- your content is organized properly, so that users can find it quickly. The larger the Web site, the more critical metadata becomes. Work done in planning and designing a quality classification pays off handsomely in the long term.
Without proper metadata, visitors will take longer to find the right content. If the site is large, it may be almost impossible for visitors to find what they're looking for. When site visitors get frustrated, they won't come back to your site, and you can lose customers or valuable contacts
XML (extensible mark-up language) is an emerging world standard for metadata. Using XML, organizations in a particular industry can agree to structure their documents in the same way (using the same layout structure and identifying the same metadata such as author name, date, and so on). People receiving these documents would be able to search them in a far more comprehensive manner.
Perhaps the best-known example is Yahoo, which focused on creating a directory classification for the Web, instead of simply depending on search engine software. Yahoo became very popular because it had professional editors selecting and classifying Web sites for its directory. Visitors to the Yahoo Web site do not get an endless listing but a selection of the best Web sites under a particular classification.
- Always keep in mind the type of person who will be looking for the content. How would they like the content classified? When carrying out an advanced search, how would they like to refine their search?
- Collect only metadata that is genuinely useful. If you ask for too much metadata, you'll slow the publishing process and make it more expensive.
- Make sure that all essential information is collected. If copyright information is needed on every document, then make sure that copyright is part of the metadata list.
- Discourage abuse of metadata. Some people will put popular keywords in their metadata just to increase the chances of their documents coming up in a search. However, this is counterproductive, as the document will not be relevant to the search in question.
- Remember that metadata should be strongly linked with advanced search. The metadata that you define becomes the parameters by which advanced search is refined.
To collect metadata, some form of document template will be required to contain all the relevant metadata fields. Examples of templates include Event Template, Technical Paper Template, and Personnel Details Template. Try to avoid having too many templates, which can be confusing. Give templates names that describe their function (instead of "Template A," for example, call it "Event Template").
Classification is a particularly important form of metadata. A Web site with poor classification is difficult to navigate, and when visitors are not able to find what they want, they will leave frustrated and not come back.
The top-level classification of your Web site expresses, in the fewest words possible, the nature of your business. Are you selling "products," "services," or "solutions?" Do you offer solutions for "home users," for "small businesses," or for "large businesses?" Fail to get these classifications correct and complete, and your Web site may become unusable. If your Web site has a large number of documents that require classification, you may need to hire professional help to get the job done.
The way you design classification has a significant impact on success of your Web site, the chances of which will be improved if you:
- Make sure that senior management at your organization is involved, particularly in design of the site's top levels.
- Design from the point of view of the people who will be using the classification. Remember that classification terms that may be understood within the organization may not be nearly as clear to customers.
- Focus on simplicity. Avoid using jargon, ambiguous words, or complex terms.
- If possible, avoid going more than five levels deep for a classification design. Remember that the more levels there are, the more clicks are required for visitors to find what they're looking for. In addition, a classification with a lot of levels is prone to error. As a general rule, aim for no more than 10 classifications at your top level. You have more flexibility at lower levels, but remember that if you have too many terms at a particular level, you risk confusing the visitor.
- Further break down any classification that would include more than 50 documents.
- Remember that the classification you create will be presented as navigation on the Web site. If possible, make sure that the classification terms at any one level—particularly the top level—are roughly the same length, or the navigation will look awkward.
- Design the classification for the entire content environment, not simply the content you wish to publish at this time. You need to accommodate documents that you may want to put on the site at a later date.
- First, do your research. What sort of content do you have right now and in what way, if any, is it classified? How are your competitors classifying content on their Web sites? Are there any industry trends emerging? Make sure you get opinions from your customers on how they would like to navigate and search through your content;
- Start at the top and design down. If you don't get the top level of the classification right, all other work is pointless. Don't hesitate to make decisions, test them, and the make changes, until you have it right.
- Get approval for the top level of classification—before you get to work on the lower levels.
- Design the lower levels. Individual departments or sections may design classification for levels that relate to their areas, but the overall design must be consistent.
- Conduct a final review—and obtain final sign-off. Classification must be properly reviewed and the appropriate sign-offs obtained before any content is published.
If content can't be found quickly, all the work and expense put into creating it in the first place will have largely gone to waste. Metadata is especially critical to the success of larger Web sites because it is the foundation on which those Web sites are built.
Many organizations already have internal classification systems. Unfortunately, these may reflect obscure organizational structures rather than something that is useful for the customer. When designing metadata, always keep in mind who will actually be using it in the end.
Classification design can become very complex, but if it is too complex it will confuse both the visitor who is searching for content and the person who is charged with classifying the content. If content is continually misclassified, the whole exercise becomes fruitless.
McGovern, Gerry.
Rosenfeld, Louis, and Peter Morville.
Van Dijck, Peter.
Gerry McGovern: www.gerrymcgovern.com/nt/class/metadata.htm
National Information Standards Organization: www.niso.org
Ricci, Christian. "Developing and Creatively Leveraging Hierarchical Metadata and Taxonomy": www.boxesandarrows.com/view/developing_and_creatively_leveraging_hierarchical_metadata_and_taxonomy