Many truths: Why semantic search matters in hiring
Posted 17-Sep -2019
When we think about HR, the first thing that comes to mind for most people, and managers in other domains, is ‘Hiring.’ In fact, even for HR professionals themselves, despite being just one of the various tasks that they have to take care of, ‘hiring’ occupies a special place. The reason is that ‘hiring’ is one area where some HR professionals truly excel courtesy the insights gained over the course of their career. For some of them, it is the dread of undertaking a task that never really delivers an outcome that makes everyone happy.
Whether it is the volume of hiring, or the quality of the candidates, hiring in modern days often proves to be a thankless job.
This is where Semantic Search steps in. Though it might not be the latest technological breakthrough in the domain, it is a fact that a lot of HR professionals are still apparently unaware of it or haven’t benefited from it while hiring.
At its core, Semantic Search is a search methodology which goes beyond the conventional Boolean search methods wherein search keywords are matched to their occurrence in the databases to come up with search results. The Boolean search method has its own strengths especially as far as simplicity, precision, and well-known skill requirements are concerned.
The Semantic Search technology offers two major value additions. The first is that it redirects the search to the candidate’s context, rather than merely focusing on what they wrote. The bottom-line is that the candidates don’t prepare the CVs for a software.
Therefore, it can analyse synonyms to extract meaning. Hence, a two-word search term can result in a 15-word search using relevant keywords. For instance, a phrase like heavy vehicle driver is likely to come up with candidates who wrote ‘operations at mines’ on their CVs.
Another example of Semantic Search’s utility is when it helps analyse things like where their best hires had worked previously, the institutes they studied at, or their past professions etc. Traditional search is unlikely to lead to such candidates, unless you intuitively added specific search terms. Hence, Semantic Search is not only capable of offering more results, but also a better pool of candidates to choose from.
Advanced Semantic Search can ‘understand’ what an applicant is trying to convey, leading to better matches, and also saving a great deal of time that is wasted in cleaning the mismatches. For instance, a simple term like ‘concept selling’ usually refers to either selling of an innovative new product or offering real-time customisations while pitching to a potential client. Semantic Search is more likely to consider both descriptions, rather than leading to a fruitless search for the precise term ‘custom selling.’
An effective Semantic Search technology for internal purposes is likely to integrate various in-house databases too. It will further allow integration with channels like a candidate’s social media profile as well.
Some of the biggest hiring platforms in the industry have either already incorporated or are in the process of incorporating Semantic Search options for recruiters who scan through their huge databases. In fact, Semantic Search performs better, even for ‘hard’ skills that while having synonyms, have a specific meaning. For instance, a search for ‘Team Leader’ might not give you better results even if you use Semantic Search method.
However, searching for a java programmer is likely to lead to candidates with mentions such as JSP, CSS, C/C++ or even Clojure, depending on the country you are located in. Being closely related to each other, and a natural progression for a person who knows Java, such results are likely to create a better search pool than conventional search. In the same way, a search for ‘android developer’ might mean a wider search net that includes those having
experience with APIs, material design or Android SDK. Especially for the recruiters, who are not well versed with the nitty-gritty of it, Semantic Search can save the day. It learns from the past search results and creates connections between terms. Hence, it keeps ‘learning’ to produce better results to a search query. It even covers for spelling errors in search terms.
Now, if that doesn’t make you happy then what will?
Co-Founder-Noble House Consulting