Table of Content

• How does Discover support fields to search jobs for candidates?
• Title
• Zipcode and City 
• State 
• Country 
• Worksite Options 
• Employment Type 
• Discipline 
• Specialty
• Certification 
• Minimum Pay Rate 
• Contract Duration 
• Skills and Keywords 
• Sorting and Display Limit Options
• When to use Engage vs Chatbot? 
• Based on the industry of jobs, which fields are recommended to use?

Legend

Recommended or Matched Jobs

How do job matching fields search jobs for candidates?

Title

To capture and run a search on the candidate's job title and jobs to find the best job match.

Below are examples illustrating the functionality of how title matching works:

• If the candidate's title = "java engineer", it will match with a job with “java” or “engineer” anywhere in the title or job description
• if the candidate's title = "RN" it will match a job which contains “RN” or "Registered Nurse" or “Nurse” anywhere in the job title or job description
  • In the case of nursing jobs, the usage of title is discouraged. Use specialty and discipline for accurate results.

Example:

	Candidates	C1	C2	C3	C4
Jobs	Title	Java Engineer	Registered Nurse	RN	Software Developer
J1	Registered Nurse
J2	ICU Nurse
J3	Junior Developer
J4	Senior Java Engineer
J5	Accountant

Zipcode and City

• Let’s take an example. Say the candidate’s zip code is 12345, and the distance is set as 50 miles in Engage or Chatbot. 
• When this candidate is searching for jobs, Discover will recommend jobs within 50 miles of the candidate’s zip code, 12345.
• If the admin user does not set distance, Discover will apply a default distance filter of 25 miles.
• Unless specified during touchpoint or node setup, the default distance unit is miles.
• Distance works only with Zipcode and City filters and not other filters. Usually, this is to recommend jobs within a certain radius to which candidates prefer commuting.

State

• It is possible that the same state or city name appears in different countries; for example, York is in the US and UK. In such situations where customers have candidates and jobs in multiple countries, they should use state, city, and country in conjunction so that job recommendations could be more targeted for specific locations.
  • Recommending a job in York, UK, to a candidate in York, US, is a poor candidate experience.
• For healthcare candidates licensed to work in multiple states, Discover can recommend jobs based on their state licensure. When searching, customers can map the candidate’s state licensure field to this field.
  • While searching jobs for candidates based on the candidate’s state licensure, avoid using city or zipcode because doing so will filter down the job recommendations to that particular zip in one state.

Example:

	Candidates	C1	C2	C3	C4
Jobs	State	Ohio	New York	California	Alaska
J1	New York
J2	Ohio
J3	Ohio
J4	Alaska

Country

• It is recommended to use country in conjunction with state or city filters for better accuracy, especially when the agency has jobs outside of the US. 
• Say the candidate’s country is set as IN, which means India. If only a country filter is used without state or city, the model could return jobs from Indiana, the US, and not India. While the model is not doing anything wrong, it will be perceived as a poor result for the candidate.

Example:

	Candidates	C1	C2	C3	C4
Jobs	Employment Type	India	United States	UK	Italy
J1	IN
J2	US
J3	USA
J4	United Kingdom

Worksite Options

• The way this filter works is that, for a candidate, it recommends the jobs that match the candidate’s worksite options.
• If the admin sets this up in Engage as a filter to be used, and If one of the candidates in the list doesn’t have data for this attribute, all jobs are recommended to them.
• If the job doesn’t have a worksite, it will be recommended to all candidates.

Example:

	Candidates	C1	C2	C3	C4	C5
Jobs	Worksite Options	Remote, Hybrid	Hybrid	Travel	On-site
J1	Remote
J2	Hybrid
J3	Onsite, Hybrid
J4	Travel

Employment Type

• Acts as a filter 
• If the value is not available in the candidate entity, all jobs are recommended to the candidate
• If the job doesn’t have employment type, it will be recommended to all candidates

Example:

	Candidates	C1	C2	C3	C4
Jobs	Employment Type	Full Time	Part-Time	Contract
J1	Full Time
J2	Part-Time
J3	Part-Time
J4	Contract
J5	Per Diem
J6

Discipline

1. If the candidate’s profile doesn’t have discipline data, they’d get all the jobs recommended.
2. This filter has intelligence built into it to decipher some industry abbreviations into actual disciplines. For example, if the candidate’s profile mentions their discipline as an RN, it will match a job where the discipline is Registered Nurse. Another example is LPN or Licensed Practical Nurse.
3. If the job doesn’t have Discipline, it will be recommended to all candidates searching for jobs.

Example:

	Candidates	C1	C2	C3	C4
Jobs	Discipline	RN	Allied Health	CNA, LPN
J1	RN
J2	Allied Health
J3	Certified Nursing Assistant
J4	Licensed Practical Nurse
J5

Specialty

1. If the candidate’s profile doesn’t have specialty data, they’d get all the jobs recommended.
2. This filter has intelligence built into it to decipher some industry abbreviations into actual specialties. For example, if the candidate’s profile mentions their specialty as “RN” it will match a job where the discipline is “Registered Nurse.” Another example is LPN or Licensed Practical Nurse.
3. If the job doesn’t have Discipline, it will be recommended to all candidates searching for jobs.

Example:

	Candidates	C1	C2	C3	C4
Jobs	Specialty	ICU, Cardiac	M/S, Cardiac	Labor
J1	Intensive Care Unit
J2	Cardiac
J3	MedSurge
J4	Labor & Delivery
J5

Certification

If the job requires multiple certifications, candidates with all the necessary certifications are recommended. Use this filter with extra caution because it could greatly narrow down the results.

Example:

		Candidates
		C1	C2	C3	C4
Jobs	Certification	ACLS	ACLS, BLS	ACLS, NRP, BLS
J1	ACLS
J2	ACLS, BLS
J3	BLS, NRP, ACLS
J4

Contract Duration

• We save the contract data in weeks in the backend wherever possible.
• Let’s take an example. Say the candidate prefers to get jobs with at least a 1-year contract duration. In this case, Discover will first convert this into weeks, which is 52 weeks. Then Discover will recommend jobs where the minimum contract duration is greater than 95% of 52 weeks, which is 49.4 weeks.

Example:

		Candidates
		C1	C2	C3	C4	C5
Jobs	Contract Duration		90 Days	2 years	4 months	6 months
J1
J2	90 Days
J3	1 Year
J4	4 Months
J5	0.25 Years

Minimum Pay Rate

• Let’s take an example. Say the candidate prefers jobs that pay at least $100 per hour. In this case, Discover will recommend any job whose minimum compensation is greater than 95% of the candidate’s desired pay rate, which is $95 or more.
  
  
• At the time of onboarding, check the default pay frequency and default currency of their jobs with the customer. This needs to be set up in the agency config during onboarding as an optional set. By default,
  • Frequency = Hourly
  • Currency = USD
    
    
• The system can also be customized to match the pay frequency and currency preferences of the candidate. In case the agency captures the candidates' and jobs' pay frequency and currency, they can further customize the search using those filters from the UI. If they don’t set it up, Discover uses the default values in the agency configs when onboarding Discover.

Example:

	Candidates	C1	C2	C3	C4	C5	C6
Jobs	Min Pay Rate (hourly)	$50	$100	$500		$300	$0
J1	$15
J2	$150
J3	$400
J4
J5	$0

Skills and Keywords

• The keywords and skills filters work in the same way. The difference comes when the jobs are ordered by relevance.
• Both the fields have keyword matching techniques, which means that at the time of the search, Discover uses a candidate’s skills or keywords to find jobs with matching keywords in the job title or the description.
• It is recommended to avoid using broad-meaning keywords, such as management. The lack of specificity in skills or keywords will result in poor results.

Example:

		Candidates
		C1	C2	C3	C4
Jobs	Keyword or Skills in JD or Title	UI design, project management	Oracle, team management	accounting	MS Excel
J1	UI design, MS Excel
J2	Oracle, JAVA
J3	Management
J4	Accounting

Sorting and Display Limit Options

How are jobs sorted for candidates and limited by a number of recommendations?

1. Candidate filters are applied on all the jobs: A list of matching jobs is created for a candidate
2. Sort Order
   • Pay Rate: Descending order of the pay rate.
     • When using an example, in healthcare, the jobs recommended to a candidate would be identical, say, ICU jobs in 5 hospitals.
     • Use this in healthcare 
   • Distance: Ascending order of the distance
   • Relevance: Fields that go towards relevance are title, skills, and keywords. The title has the highest weightage; skills have 50% of the weightage of the title, and keywords have 50% of the weightage of skills.
     • Discover will internally build an ordered list matching jobs for the candidate. In this list, the first four jobs will have title relevance, the next two jobs will have skill relevance, and the next one job will have keyword relevance. It will continue to do so until all the matching jobs are ordered.
     • Discover takes the candidate’s title, skills, and keywords and looks for those in the job title and job description while determining the relevance.

3. Number of Recommendations: The final list of recommended jobs displayed to the candidate is governed by the limit set by the customer in Engage or Chatbot. 
   For example, if the number of recommendations is set to 5 and Sort Order is by Pay Rate, it will show the top 5 jobs by Pay Rate in descending order (highest pay rate job at the top).
   
   
4. Data Quality: A key tenet of Discover’s success is the quality and completeness of customer data. While the customers have the flexibility to use any filters they want in Engage or Chatbot, they should ensure they have good data on the job order entity and candidate entities so that Discover’s matches are highly relevant.
   • Data Debugger is a great tool for inspecting the quality of the jobs and candidate data and helping customers understand the gaps in the ATS data. You can choose any entity, provide an ID, and inspect the fields of that entity to ensure the information the customer is sharing is accurate and, in case it isn’t, enables you to provide the right feedback. 
   • In many cases, the field isn’t synced into Sense. Use the simple feed sync tool to sync the fields into Sense before onboarding the customer.

When to use Engage vs Chatbot?

• Engage works best when the ATS job and candidate entity have good candidate data to produce better job filter results. Many customers struggle with poor match quality because their ATS doesn’t have good job and candidate data.
  
  
• In situations where candidate data is unavailable or up to date, recommend using Chatbot. Chatbots can collect the latest candidate information and search for jobs based on that. This reduces the reliance on the ATS data to produce higher-quality matches.

Based on the industry of jobs, which fields are recommended to use?