What
In Mercury Search there is a rating system which shows in the UI results as Match Rating. It's a dynamic value that updates whenever a Search query runs. This feature is a great way of identifying Candidates who best match a certain criteria. Boost can be utilised as part of this value mapping to streamline results even further. The key benefit of understanding these values and how they impact search, is that users can rest assured that only relevant Candidates will be displayed, therefore saving time in manually searching for the ideal Candidate to fit a certain role.
In simple terms:
Match Rating is: "If a Candidate matches on more search terms than another Candidate, they will be matched higher”.
NOTE: Match Rating is the default value in which search results are shown (highest > lowest).
Boost is: "A value will affect the priority of individual Search term results".
NOTE: Boost cannot be used in the Boolean search mode.
Match Rating
Match Rating is based on six search criteria. The top Candidates will be those most relevant, and that match the search criteria more closely across multiple scoring factors:
- Term Frequency
- How often a search term appears. More frequent = higher score.
- Inverse Document Frequency
- Commonality or rarity of a term. Rare term = more significant = higher score.
- Field Length Normalization
- Normalised score based field length. Search term in shorter field = more significant = higher score.
- Coordination Factor
- The number of matched query terms. More matched search terms = higher score.
- Boosting
- Use Boost to give higher importance to a filter field. Higher Boost number out of 10 = higher score.
- Proximity
- For phrase searches, the proximity of terms can affect the score. Search terms appearing close together = higher score.
Where
- Match Rating is found with each Candidate result, in both Tile and List view.
- Boost is found within a majority of search term filters.
Match Rating
Tile view
Match Rating is visible by default in Tile view.
NOTE: The colour system is used to understand at a glance how closely a Candidate matches the search terms:
- Blue: Highest
- Green: High to Very High
- Yellow: Medium to High
- Orange: Medium to Low
- Red: Lowest
List view
List view has the advantage of giving an exact rating in terms of percentages (0=100%). If Match Rating is not visible, add the column by selecting Match from the drop-down menu in the top left-hand corner:
Boost
Along with using filters to narrow down the best Candidates, use Boost to highlight what filters are most important in your search. Not every search criteria will hold the same weighting, so use Boost to your advantage!
Boost is currently available for the following filters:
- Job Title
- Tags
- Location (Free Text)
- Employer
Example: A Recruiter is looking for Candidates whose current Job Titles contain either 'developer' or 'tester'. Both search terms get the default Boost value of 5.
- The top two results are Chris Matthews, a Tester, and Ravi Parmar, a Developer.
- Chris has the higher match rating out of the two.
- If the Boost score for 'developer' is increased to 7, note how this impacts the results (sorted by Match Rating by default).
- Ravi Parmar is now the top Candidate, whereas Chris Matthews is now 9th in the results list
END