This week I’m pleased to present a guest post written by my colleague and researcher on the InnovateUK project Jon Chamberlain. Jon’s been doing some interesting work on analyzing Boolean strings and visualizing them using spider diagrams. Over to you Jon!
Visualising how recruiters search
Recruiters are a hard working bunch, estimated to put in a 55-hour week on average, and yet 76% of businesses report feeling they don’t get value for money from recruitment professionals. Clearly if you’re a recruiter you can’t work harder, so you need to work smarter to stand out in a growing and highly competitive field.
Why visualise your search queries?
Recruiters use fine-tuned search strings to find candidates, targeting candidate qualifications, previous employment and/or location. Search operators such as “AND”, “OR” or “NOT” can be used to include and exclude certain keywords, ending up with a complex query like the example on the left.
It can be hard to understand how one query is different from another, other than by looking at the results it brings back. Perhaps one of your queries worked very well but others are not so good. Why?
To find out, this prototype can visualize a search query using a set of common dimensions such as query length, number of operators and keywords used.
Learning from other recruiters’ search queries
The inspiration for this was an experiment set up by Glen Cathey on his Boolean Black Belt website where he asked his readers to contribute a search query solution to a candidate sourcing problem. The aggregate of 46 responses can be visualised below:
Glen’s experiment showed that the contributors used plenty of keywords and key phrases, combined mainly using the “OR” operator (although in many cases “AND” is the default operator when no operator is explicitly defined so that count may be slightly misleading). Operators such “NOT”, “NEAR” and wildcards were rarely used.
Glen was good enough to provide examples of how he personally would solve his candidate search problem, which can be visualized as follows:
The first thing you notice between the 2 visualizations is that Glen’s answer is much bigger in area, meaning that his query was longer and more complex than most recruiters’ answers. The visualization shows that Glen used a wider range of operators in his answer, in particular wildcards that allow for variations of words and the “NOT” operator that would remove irrelevant candidates from his results set.
The V-shaped search
Another source of recruiter queries can be found on the Boolean Strings Ning network, where recruiters were invited to add their “best” search query, although in this case there were no criteria for the query construction. The aggregate of 41 queries produces a very distinct V shape:
This visualization shows that the recruiters here focused mainly on using keywords and the “OR” operator. In comparison to the queries from Glen’s experiment these recruiters are not using Boolean search operators to the same extent, although this may be because there was no specific sourcing problem defined.
Given the highly competitive nature of the sourcing industry, recruiters would benefit from learning more about how to fine tune their search techniques to find candidates – work smarter not harder and let the technology do the legwork. There are some excellent free query building tools available on the web like Recruit’em or SocialTalent to get you started, as well as specialist training in Boolean searching for recruiters and communities like Undercover Recruiter offering regular posts on the latest recruiting tools and techniques.
Share your experience of searching for sourcing
To find out more about how recruiters search, UXlabs is running a larger survey to discover how queries are constructed and managed. If you are a recruiter please participate (anonymously) in the survey. You will be given access to a pre-publication copy of the report, as well as a chance to win $100.
By now you’re probably wondering how your search queries will look when visualized. Visit http://jonchamberlain.com/ssa.php and enter a query into the box on the left (or run the example query).
This article was written by Jon Chamberlain on 22 July 2015, please get in contact (firstname.lastname@example.org) if you wish to reproduce any part of the this work or would like more information.