Wasting Time With Online Job Search
Our latest research report shows that most of the results obtained after keyword or filter search on online job portals produce little or poor fit with what the job-seeker intended to find. Who knew? Ludmila, armed with research, delves into this topic and reports her findings.
What's going on?
Per our findings, 66% of the results generated via online search on job boards were for jobs that our respondents would never apply for. No doubt, this impacts the sourcing process: for the job-seekers, it means that they spend more than half of their time filtering through irrelevant job suggestions; for the recruiters this means, that more than half of their jobs are not targeting the rights audience.
Seriously, what is going on? How did we get these results and what are the possible explanations of the person-job mismatch? (our futile attempt at starting a hashtag #personjobmismatch)
We observed 35 volunteers using a made-up job board to find employment in Denmark. Our typical respondent was a foreign-born active job-seeker between 25-35 with a university degree and work experience.
Respondents were asked to use keywords or filters to search for their desired job and analyse the search results – would they apply for this job? Why or why not? After hearing their responses, the researcher would suggest them to modify search results in order to get more accurate results. The modified search results would be analysed in the same way. Altogether, we analysed 249 results, an average of 7 job results per respondent.
Fit or no fit?
Over a half of respondents used the English version of a popular Danish job site, which brings up the possibility of the low person-job fit being due to language discrepancies. However, our evidence suggests that language is not the main reason for the low fit.
The main reason why respondents found the results unfitting was actually the mismatch with their qualifications. Most results were categorized as being “completely off” meaning that they didn’t match their educational or professional background. The most striking examples of these were “nurse” after looking for the keyword “management consultant” or the job vacancy for a “universal banker” after looking for “non-profit arts.”
Another type of qualification mismatch was due to the seniority requirement of some positions. These were, however, matching the professional and educational background of the job-seeker, just not at the right level.
The second most popular reason for respondents rejecting a job result for low fit was that the job required experience – either in the sense of years of experience or experience from a particular industry or/and a particular tool. Language was the third reason why respondents would not apply for a job.
Keyword or filters: what generates better results?
The ultimate question, asked on a daily basis since the beginning of time (at least we think); most respondents used keywords rather than filters to conduct their search. The keywords were usually related to job titles.
The fact that almost all keywords used were related to the job title suggests that using an unsuitable job title can account for the low fit of the results. However, as our research showed, there is no significant relationship between using keywords or filters and generating poor result – both keyword searches and filter searches generated approx. the same proportion of results reported with “no fit” or “poor fit”. This also supports the argument that language did not play a major role in generating low person-job fit, as filters did not require respondents to use exact terms in Danish and yet generated the same percentage of low fit results.
It’s the profession, stupid!
One can hypothesize that low fit results of job search have something to do with the professional area one works in. The more “generalist” type of jobs, such as sales and marketing people, management and operations people, lawyers and analysts can see themselves in several different positions within an organization. That’s why when they look for jobs online using keywords or filters, they receive a larger and broader spectrum of suggestions compared to a person looking for a specialist profession, like doctors, teachers, engineers, etc.
Our results show this is partly true. Complete fit ism ore likely to be achieved for people in more specialized professions, such as engineering. However, “no fit” of search results has been observed across all industries and professions.
Why do we think we’re fit for a job?
When the job seekers found a vacancy interesting, we asked them why. The majority of people answered that it’s because it matches their skills and knowledge of the area. They formed assumptions of skills match based on having performed similar tasks in the past. Knowledge match would be achieved in their eyes if they either had an educational background in the job or/and had worked in the same or similar industry. People were less likely to apply for jobs where they only matched the required skills or the required knowledge.
"There is no significant relationship between using keywords or filters and generating poor results while searching for a job."
To maximize the efficiency of searching for a job online, you should make it as accurate as possible. It is useful to do research on job titles relevant for your professional background first and then test out different words and their combinations to target your profession.
Inserting specific skills or tools you can use (e.g. “CRM” or “ServiceNow”) instead of practically guessing relevant job titles produces more useful results.
Don’t ignore jobs for which you have the relevant skills or knowledge but not both. Applying for jobs that you have both knowledge and skills in is what everyone does. This means that coming into a job with different skills or knowledge is something that can make you stand out more – and that way actually raise your chances of getting that job!
"Coming into a job with different skills or knowledge is something that can make you stand out more."
Recruiters and hiring managers
Most job-seekers search for jobs using job titles as their keywords. The more standardized job titles you use, the greater the chance of your job ad being found by relevant candidates.
Experience was the second most popular factor for deciding not to apply for a job, i.e. the number of years and/or industry relevant experience. Make sure your expectations are realistic, otherwise you risk losing good candidates just because your base criteria were too high.
Most job seekers apply for jobs where they match both the skills and knowledge requirements. If your demands in both areas are too specific, you may miss out on good candidates that could transfer relevant skills and knowledge to the job.
Online job search is, in practice, time consuming and dependent of the user’s awareness of all suitable job titles on one hand, and recruiter’s awareness of standardized terms and realistic demands set on candidates on the other.
With currently available technology and limited human capacity it is only possible to sort through all relevant information by applying pre-determined criteria, such as keywords and filters. These do not, however, generate satisfactory results because the principles of a “match” are much more sophisticated.
Machine learning algorithms analysing big data offer an opportunity to generate person-job matching suggestions with humans only required for the initial data input. They also enable ranking of suggestions allowing the user to decide – in this case – if he/she wants to only apply for positions where both skills and knowledge match is achieved or also for those, where only one of them is achieved.
If you would like to see the full research report, contribute with your comments or ask questions about this research, please email firstname.lastname@example.org.