A good label complements the abilities of Google search to help users find relevant webpages. Labels represent aspects of webpages that are hard to capture with normal query terms in a search. We have identified a few categories of labels to consider:
- Audience: Any term you can use to complete the following sentence: 'This page is written for ____'. Labels such as 'for_students', 'for_patients', 'for_kids', 'for_lawyers', etc. are very useful as the intended audience is hard to pick up with search query terms. We recommend the naming convention 'for_*' for these labels.
- Source Type: Any term you can use to complete the following sentence: 'The author of this page can be categorized as ____'. Are your subscribers looking for material from corporations, universities, government, nonprofits, etc.? We recommend using the name of the source category ('government') as the label as the source of a document is veryimoprtant in sifting through results.
- Document Type: Any term you can use to complete the following sentence: 'The contents of this page and others like it are called ____'. Many webpages fall into a category of document types (images, maps, catalogs, biographies, interviews, movie_reviews, news, forums, demos, downloads, specifications, product_manuals). These are all excellent labels because it is hard to restrict a search to a specific document type strictly by using search terms. By convention, we use the plural noun for document type labels ('biographies' instead of 'biography').
- Subject: Any term you can use to complete the following sentence: 'This page is about ____'. We don't recommend subject labels because query terms are usually sufficient to find relevant documents. However, in some cases even when the subject is prevalent on the page, the subject label is useful in identifying good quality pages. For example, there are a lot of pages on the web with the word 'review', very few of which contain actual reviews (most are inviting users to write reviews). In this case, it makes sense to have the label 'review'.
- Label names should be all lowercase.
- Punctuation and conjunctions ('and', 'or') should be reduced to '_'. The phrase 'warnings and recalls' becomes the label name 'warnings_recalls' and the phrase 'tests/procedures' becomes the label name 'tests_procedures', and so forth.
- Label names should be as short as possible. On the search results page, labels may be truncated to save space, particularly if a URL has many labels.
Finally, we have a naming convention in terms of punctuation. Consistent punctuation can help ensure that you and others will have the happy coincidence of using the same labels for the same kinds of documents.