Problem: Too many, and irrelevant results. Solution: Facets!

Faceted Search

Finding just the right documents in a huge corpus of diverse text can be tricky, if all you got is a text field for entering arbitrary search words. A bit of assistance is usually appreciated. Facets is one way of providing this!

Assisted Search

Searching is only a success, if you find something relevant – perhaps the item you were actually looking for or perhaps several articles relevant to your inquiry. Oh, and only if the search process was easy and quick. 

When searching in large corpuses of contents with which one is not intimate familiar, it is often quite a challenge to figure out just which search words to use and in which combination should be entered in order to get the most relevant results, and not all sorts of other irrelevant material in the result. 

Very often, the visitor can benefit greatly with some assistance in the search process. Such assisted search can be achieved in several ways. One very common method is to provide Auto-completion (Type Ahead) in the search box and Did you Mean feedback along with the search result. Another very common way to assist the search is to apply facets.

Forward Search provides all these search assistance tools as easy-to-use features ready to utilize Out-of-the-box

Introducing facets

Facets is a term often used in the search solution context, and covers properties of the searched content that is not the main, unstructured text, but instead a number of meta data describing the content in sets of finite values which could be keywords, tags, numerical figures, dates and similar properties that classifies each item. Each set of similar values is referred to as a “facet” of the contents, and these facets are often very useful for providing the assisted search we are trying to achieve. 

Using facets to assist the search 

Facets are typically applied by presenting them as part of the search window in the form of drop-down selection options, checkboxes and range-selectors. When the query is executed, these facets are gathered and added to the query, filtering the result further, and hopefully returning a more relevant result set. The chosen facets, as well as those not selected, are most often displayed with the result, in order to allow quick further filtering, or indeed removing some of the applied filtering selections again, in the case the search became too narrow. This is often referred to as “Search Refinement“.

A notable difference between the concept of “facets” and simple keywords or categories, is that we often associate a frequency on each facet value – a frequency of documents carrying this facet value within the current search result. These frequencies are very often shown as part of the search window both initially and in the search refinement context, to provide further feedback to the user about just how many documents there will be carrying a particular facet value. 

See the examples below on how this may typically look.

Facet search examples

Search Driven Navigation

Facets can be applied to the contents solution in more profound ways, more or less eliminating the need for the classical search field. The facets can be presented as navigational options along with, and seamlessly integrated with the standard website navigation system, as sub-menus and multiple choice options on overview pages. We often refer to this use of facets as “Search Driven Navigation”.

Facet setup and administration

Forward Search is perfect for facet setup management. Our extended Custom Fields are the perfect vehicle for facets, allowing a number of relevant parameters to be applied to each facet in the index. We even provide a specialized analyzer and tokenizer particularly useful for facets, and it is not surprisingly called the “Facet Analyzer” when selecting it in the Forward Search Administration client.

To introduce a new facet in an index, 3 steps are needed.

Firstly, the value of the facet must be exposed on the document to be indexed – as an html meta-tag or a specially labeled html contents element on the body. Or, alternately, it can be extracted, calculated or looked up and added the document at index time, utilizing the Finalizer API

Secondly, the facet field must be set up for the index in question – more precisely for the contents source where these values are exposed. This setup is done in a matter of minutes in the Forward Search Administration client.

Finally, you need to re-index the site, in order to get the new facet values extracted and added to the index. This is easily done from the Forward Search Administration client - just click the “Recycle” button on the index in question. Once the re-index is done, your facet is ready to use. 

There ARE other means of adding new facet fields to an already build index, but that involves more complex steps and is covered in our technical user guides.