From our Depsters May 23, 2018
How machine learning helps getting your jobs in front of the right candidates
Google’s Cloud Job Discovery is disrupting the online job industry – or so, it says. Their goal? To get the right people to the right jobs and vice versa. But what really makes Google’s solution different than the current job boards already in the market?
Well, the biggest issue with companies finding the right people for their jobs and job seekers finding employers that fit is that most of the time they simply can’t find each other. The world wide web already brings supply and demand closer, but the total amount of vacancies is overwhelming and most candidates can’t see the wood for the trees.
But of course candidates don’t just look at all the vacancies available. They search for jobs close to home that require certain expertises and fit their area of interest. Then, when job boards receive a search query they match it with their vacancy index based on keywords. This seems like a good way to match job seekers with vacancies, because if the keywords they search for are in the job description, then it must be relevant. Right?
Wrong. Obviously lots of search results are somewhat relevant, but more often than not the search results show a mismatch. A common result can be found in the example below.
This is annoying. For job seekers, because they get tired of wasting time on poorly matching vacancies. But also for employers, because they invest a lot of time and money in finding the right people for their jobs by advertising on job boards and by optimizing their job descriptions for certain keywords.
Luckily, Google now offers Cloud Job Discovery that promises to optimize the process of job seeking.
What does Google’s Cloud Job Discovery do?
Cloud Job Discovery uses the machine learning capabilities that are also used to match search queries in Google to show users the best results for their search intention. It’s a plug & play cloud service that can be used for all kinds of job properties, such as job boards, career sites, staffing agencies etc. Note that getting the best match for a search term or for a searcher’s intention can be two entirely different things.
So, how does that work in real life?
Google uses machine learning to discover what job seekers really mean with their search query. This means they don’t just look at the keywords that are typed in at that moment, but they use the knowledge the machine acquired through all other search queries as well. In other words, the model continues to improve over time, since the amount of data the machines learns from keeps increasing.
Besides that, the model is pre-trained to keep in mind the use of slang in search queries, job title classification, occupation based semantics, location you are searching from, synonyms, the detection of keywords used for skills etc. Besides, Google takes into account the preferred commute time and mode of transport of the searcher.
As a result of all this knowledge, the Cloud Job Discovery search engine should return far more relevant job postings to job seekers. That will then lead to less time spent searching and a far better experience for job seekers. At the same time, it gives job posters more value for their money because the people looking at a job posting will be more likely to actually apply.
How Google’s Cloud Job Discovery would match a search query with a vacancy is as follows:
First, the content on the job site using the Job Discovery functionalities is inserted into the Google cloud platform. The Job Discovery then starts normalizing job titles into commonly searched for titles and starts mapping all content to pre-trained data models that codifies the content and makes it “understandable” for the search algorithm. The system continues to recognize keywords in search queries like they are intended, such as skills, locations, benefits and certification.
As a result of this algorithm the search results will not only show the best match to someone’s search query, but also similar roles nearby even though they might be written with totally different keywords.
So, it doesn’t matter if you are trying to find your dream job or if you want to get your jobs in front of the right candidates. Machine learning might just bring you a lot closer to your goal. Check out Google’s Cloud Job Discovery webpage to learn more or to request access.