by Addie Van Salisbury on Feb 18, 2021
One of the main focuses for Aspen Discovery is to connect your patrons with the title and item they are most likely looking for. We have all had a frustrating experience where our search for ‘Harry Potter’ returns a magic cookbook or a film score. We want to ensure patrons are aware they have access to the materials they are most interested in and that they easily can find those items.
This week I want to cover some of the many ways Aspen handles relevancy and how you can manually control your catalogs' relevancy. All of the factors below are considered when Aspen Discovery is determining the relevancy and choosing what results to display to patrons when they search your catalog.
Keyword Search Relevancy
Aspen uses specific ‘boosting’ criteria when looking at the different types of searches that you or your patrons are doing in your catalog search. Our code will examine the input of a search and depending on where in the record that search matches it will apply a ‘boost’ which is a number associated with the records relevancy. The higher a boost is, the more applicable, or relevant, a title is. When using Keyword search, if your input exactly matches the title of a record, then that record will get a huge boost (10000).
If the search term matches the beginning of the title it will get a boost (8000). As the title match is less exact, the relevancy will decrease accordingly.
If your search in keyword exactly matches an author field, then we get a boost of 2000. If the search matches the beginning of the author, we get a boost of 1000 and if the search hits other parts of the author it does take that into account for relevancy but much smaller (from 50-200 depending on what parts of the author it hits).
Aspen will also check to see if the input in the keyword search matches the barcode, issn, isbn, and call numbers. For the call number, Aspen will boost an exact match on call number with a boost of 400 and if the call number only matches the beginning it will boost it with a boost of 300.
If the input you have in the keyword search matches the series information, it will boost the record if it's series exactly matches with a boost of 1500. If it matches on part of a series name it will boost some but less (30-1500).
Finally, Aspen will check to see if the search you put into keyword lives in the table of contents, subject, topic, geographic region, genre, era, keywords, description, target audience, and literary form (nonfiction/fiction).
Circulation
Aspen Discovery also considers the number of checkouts on items and titles. As patrons use a title more, Aspen Discovery will increase the relevancy of that title. The system does this because most likely if a title is very popular and being checked out often, it is a title other patrons would be expecting to see when doing a search. Depending on the ILS, we will take into account how recently the title was checked out too. For example, Sierra will store checkouts this year, last year, and all time. This is useful because if something was popular a few years ago it won't be shown as highly as something popular now. We also boost based on the number of holds for a title if that data is available. This allows us to show something that a lot of people want to check out right now higher in the results. For some eContent sources (like OverDrive) we consider the number of checkouts as well.
Number of Holdings
In the Catalog / Grouped Works section of the Administration menu, if you click into your specific ‘Grouped Work Display’ settings there is a checkbox option that says 'Apply Number Of Holdings Boost'. We generally recommend turning this on, as this will add boosting or relevancy to title records that have a lot of items attached. The thought here is that you are most likely buying a lot of copies of popular titles but maybe not as many of the unpopular titles. If a title record has 20 items attached it may show higher in search results than a title with one item attached.
Manipulating Relevancy
A library can also manipulate relevancy for their specific patrons and their needs. You can actually self-assign boosting to your formats. The scale for this is low, medium, high, and very high. low being less relevant and very high being very relevant. For example, if a library wants to make sure Books or DVDs are showing higher in relevancy, therefore higher in a patron’s search results than Microfilm then a library can assign a Book format relevancy of very high and microfilm low. Feel free to play around with the boosting scale!
In order to adjust the formats’ boosting you will go into ILS Integration -> Indexing Profiles -> Click into your profile and scroll down to ‘Format Information’. In the ‘Format Map’ there is a column called ‘Format Boost’. This is where you input your boosting number!
Read more by Addie Van Salisbury