Integrating a search engine into a website can be as easy as a single click (E.g. using a search plugin for a WordPress blog) or so complicated it requires months of development effort (E.g. for customising an existing search engine for your own needs). The various options available can also vary vastly in price. The terms used however, are similar regardless of the solution.
String Matching
Match the search terms with part of the content item E.g. a product name or a page’s title
Fuzzy Search
Programmatically searching for variations of the search terms – it should pick up most typos and use “fuzziness” algorithms to find approximate matches. Only the words that are used in the content items for the site are used when trying to find alternatives. This prevents suggesting alternatives with 0 results.
Attribute Search
Searching through a content item’s attributes E.g. product colour, page size etc.
Category Filters
Filter search results by category E.g. restrict search results to only those tagged as in the “recipes” category
Facetted Attribute Filters
Filter search results by selected a desirable attribute value. Can be used for multiple attributes simultaneously. Can be used to allow multiple values to be selected for a single attribute. Also known as guided navigation
Field “Boost”
If a field is identified as more important than others (E.g. content item name), then a “boost” factor is associated to it. When search terms are identified as being in a boosted field, then their relevency ranking is boosted.
Analytics Integration
Search queries are recorded in an analytics system (E.g. Google Analytics) to identify top searches, search frequency etc.
Document Search
Search through the content of documents (E.g. PDFs) in addition to content items
Predictive Search
Use previous search queries to “predict” what is being searched for – can indicate number of results/products for a particular search term
Hit Highlighting
Identify where the search term is within the search results and apply a background colour.
Sorting by Relevancy
Ordering search results by how relevant they are to the search term. E.g. If the first word of the search term appears several times in a content item’s name then it is more “relevant” than if the last word of the search term appears once in a content item footnote. This capability can be combined with Searchandising functionality (see below) to ensure
Searchandising
The ability to tweak how search results are displayed based on business rules, promotions, business intelligence etc. Able to manage “relevancy” when results are equally relevant to search terms. This often requires extra data to be stored against the content items E.g. sales data so that more popular items appear higher up the search results