The objective was to turbocharge the job search process by providing a great set of relevant jobs for candidates and save the recruiter hours matching the right candidate to their next job.
How to improve job searching for candidates and job matching for recruiters?
360 Systems has been working with recruitment businesses for over two decades, creating websites, vendor management applications, and online pay document and timesheet systems.
In preparation for a client's quarterly competitor analysis review, we reviewed the status of their recruitment website to make recommendations. Search matching for both the candidate and recruiter was a key focus area for the review. The website was already utilising vector search and returning accurate search results, but it wasn't able to remove false positives or include jobs that with candidate wasn't specifically searching for. This is where the job title classification system could help and turbocharge the results. What I mean by this is:
- Exclude weak matches: Remove results that are clearly not right so the candidate has confidence in the search and can focus on the strong matches. For example, a candidate searches for 'Deliver Driver' and results include jobs with the keywords 'Deliver' and 'Driver' as well as 'Deliver Driver'. Lots of jobs have the words deliver, delivering, drive etc. in the title and description which can result in false positive matches. We want to remove these as many of the jobs would not be relevant.
- Include similar jobs: With the same example, searching for 'Deliver Driver' would not return 'Courier' jobs, as the keywords searched will not match job titles without those keywords.
This resulted in poor matches and recruiters having to add additional keywords to their adverts to get them into the search results.
We felt that having a job title classification or taxonomy to categorise job titles and link them to other like roles would dramatically improve the search experience.
- It could help job seekers find more relevant jobs. By organising jobs into categories based on their skills, keywords, and other relevant criteria, taxonomy-based job search can help job seekers find jobs that are a better fit for their skills and interests.
- It could help job seekers find jobs that they might not have found otherwise. By taking into account the relationships between different terms, taxonomy-based job search can help identify jobs that are relevant even if they don't use the exact keywords that the job seeker entered.
- It could help job seekers save time and effort in their job search. By providing a more structured way to search for jobs, taxonomy-based job search can help job seekers find the jobs they are looking for more quickly and easily.
- It could help job seekers find jobs that are a better fit for their skills and interests. By taking into account the job seeker's past search history and other factors, taxonomy-based job search can suggest jobs that they are likely to be interested in.
- It could help job seekers find jobs that are more likely to lead to a job offer. By providing more relevant results, taxonomy-based job searches can help job seekers increase their chances of finding a job that is a good fit for them.
What is a job title classification system?
A Job title classification system is the process of assigning a job title to a predefined position. Adding hierarchy to the classification system allows job titles to be allocated to specific levels. The levels in the hierarchy are typically based on:
- Family: Typically includes operations, marketing, sales, human resources, finance, research and development, sales, customer support, shared services, engineering, information technology, and so on.
- Industry: Group of companies that are related based on their primary business activities. For example manufacturing or services.
- Skills & levels of complexity: The skills required to complete the specified job role.
- Duties: The typical duties for the specified job role.
- Responsibilities: The key responsibilities for the specified job role.
- Experience: A worker may require certain levels of experience in order to complete more senior roles.
- Entry routes & qualifications required: Does this job role require specific qualifications or licences to enable them to complete the role?
Designing a job title classification system
We initiated a project and utilised the ONS Standard Occupational Classification (SOC 2000) as the foundation list for the job title classification system. We understood that many of the job titles were obsolete and it didn't always contain newer job titles or a spread of 'Also known As'. We therefore used a combination of human review and AI to highlight obsolete and create new job titles, while also identifying words typically used by job seekers and recruiters when the standard job title was not the only option, for example, 'Deliver Driver' and 'Courier' which are closely related and potentially an interchangeable role for a candidate.
In parallel with the administration above a new Microsoft Azure microserviced application was created to manage to configuration of job title classifications while providing a matching and autocomplete service for identifying the best job title for a given input. All jobs were classified using this process and the search keywords input on the website was enhanced to use the autocomple. The search algorithms would then use the hieratical classification scheme to prioritise (boost) certain jobs and include others that wouldn't typically be matched. It would also remove false positives to reduce the matches while making those that are returned more accurate.
Benefits of job title classification in recruitment search
The solution created was utilised for the primary objective of 'improve job searching for candidates and job matching for recruiters' but also provided other benefits to the business:
Taxonomy-based job search
Candidates now search for jobs based on the classification and not just the keywords. This allows other job roles that are closely linked to be included in the suggested matched jobs providing more options for the candidate. We also found that the search results are more relevant by excluding jobs that would have previously been a match to the searched keyword but would not to the context of the role.
Candidate search matrix
The client categorised their registered candidates' CV data to give them each a classification job title. This information would then be used to perform searches for available candidates along with traditional keyword searches to filter and find suitable candidates.
Business Intelligence
The goal was always to improve business reporting and intelligence by capturing what was being searched for. The business already had a mechanism to capture what keywords were being used when candidates searched for jobs but it wasn't always easy to get the full picture due to the keyword input being free text and allowing for misspelling and word variations. The introduction of the job title autocomplete still allowed candidates to search any keyword but would hint at suggested job title classifications to use.
The introduction of the job title autocomplete meant that most job seekers selected the suggested job title and therefore the reporting was more accurate as to which job titles were being sought in each of their locations empowering them to make decisions around their focus in that location.
Salary Trend Data
The chairman requested a new Salary Trend service to promote the business's website and not to use external data but use their own on a rolling 2-year. As all the job data had been classified with the new job title classification system it made this both achievable and unique in that it used up-to-date real published job data to give salary data for each job title.
Conclusion
The job title classification system solution developed by 360 System now powers both candidate job searching and recruiter candidate searching, along with the salary trend data pages on their website.
The accuracy of results has provided both superior matching and remove poor matches, saving time and effort, while providing a more structured way to search.
Our client now regularly reviews how candidates use the job search and adds new job titles or 'Also Know As' to further enhance the user experience. Their recruiters' feedback is captured via a feedback system to monitor candidate matching to ensure the best results for the business.
If you have any questions or would like to discuss how a job title classification system might benefit your organisation then please don't hesitate to get on contact.
Richard Evans, Projects Director
