Results and Ranking
Why results are scored and ordered.
Results and Ranking
How Beesla Scores and Ranks Candidates
When you run a search in Beesla, candidates are not returned as a raw list. Each result is evaluated, scored, and ranked based on how well it matches the role you described.
This page explains how results and ranking work so you can review candidates confidently and make faster decisions.
How Ranking Works
Beesla evaluates every candidate returned by a search against your role description.
Ranking is based on:
- Relevant experience
- Role alignment and seniority
- Career history and context
- Signals found during the search
The result is an ordered list where the strongest matches appear first.
Ranking is role-specific.
The same candidate may rank differently for different roles.
Understanding the Score
Each candidate includes a numeric relevance score.
The score represents how closely the candidate matches the role, relative to other candidates in the same search.
Important things to know:
- Scores are comparative, not absolute
- A high score means better fit within that search
- Scores should guide prioritization, not replace judgment
Use scores to decide where to start your review.
Reading the Explanation
Alongside the score, Beesla provides a short written explanation.
The explanation highlights:
- Why the candidate was included
- Which aspects of their background match the role
- Any notable strengths or gaps
This context helps you quickly confirm whether the ranking makes sense.
If explanations consistently feel off, refining the prompt usually improves results.
Why Fewer Results Are Better
Beesla is designed to produce shortlists, not massive result sets.
A good search typically returns:
- A manageable number of candidates
- Clear separation between strong and weak fits
- Candidates you would realistically consider contacting
If you see too many marginal candidates, tightening the prompt is more effective than filtering manually.
How Ranking Affects Cost
Each candidate returned by a search costs tokens.
Better prompts:
- Reduce unnecessary results
- Lower token usage
- Improve shortlist quality
Ranking helps you spend time reviewing strong candidates instead of sifting through noise.
Common Misunderstandings
“A low score means the candidate is bad.”
It only means they are a weaker fit for this role.
“Scores should be compared across searches.”
They should not. Scores are search-specific.
“I should review every result.”
Start at the top. You rarely need to review everything.
When to Rerun a Search
Rerun a search when:
- Results miss a key requirement
- Seniority is consistently off
- The role scope changed
Before rerunning, adjust the prompt to avoid repeating the same results and token usage.
What’s Next
To see how your searches evolve over time: