Four Analytics tabs answer four questions. Velocity: where is the pipeline slow? Quality: is the bar set right, and does the AI's recommendation line up with what happened? Experience: are candidates hearing back, and how fast? Roles: which jobs are working and which have stalled?
All four use the date range, company and job you pick in the filter row at the top of Analytics, except where noted below. See Dashboard and analytics overview.
Velocity: time in each stage
The figures at the top are Median time to hire, Median first response, the Slowest stage and how many Completed stage moves happened in the range.
Time in each stage shows the median number of days candidates spent in each stage, from entering it to leaving it. Each row says how many completed moves it is based on, and how many candidates are still here. Only completed moves count: someone who has been in Interview for three days so far is not counted yet. Where a stage has five or more completed moves, a paler band behind the bar shows the spread, from the 25th to the 75th percentile. With fewer moves a spread would only describe a handful of people, so it is left out.
Where the funnel leaks is the hiring funnel again: the furthest stage each candidate ever reached, and the share that continued at each step. A low share at one step is where candidates are lost.
Time to hire runs from the application to the moment someone moves the candidate into the Hired stage.
Quality: is the bar set right?
The figures at the top: how many applications were Scored, the Average fit, the share that Clear the bar (a fit score of 70 or more) and Proceed → interview, the share of Strong match candidates who reached Interview or beyond.
The charts:
- Fit score distribution groups scored candidates into bands: 0–39, 40–54, 55–69, 70–84 and 85–100. The bands from 70 up are the qualified pool. With one job selected, click a band to see those candidates.
- Where the pool is strong and weak shows the pool's average on each scoring axis, such as required skills, experience and communication, with the weakest highlighted.
- Recommendation mix shows how many candidates the AI rated Low match, Caution and Strong match.
- Risk mix shows the spread of risk levels. Everything right of the line (high and critical) deserves a human look.
- Must-have skill coverage shows, for one job, how much of the pool shows each required skill. See Dashboard and analytics overview.
- Screening knockouts appears when your screening questions have flagged candidates with a knockout answer. It lists each question with how many answered it, how many it flagged for review and the rate. A question that flags most of your applicants may be filtering harder than you meant. See Screening questions and knockouts.
Recommendation versus outcome
This grid sets what the AI recommended against the furthest stage each of those candidates actually reached. Read it a row at a time: "of the candidates the AI called Strong match, how far did they get?"
Experience: are candidates hearing back?
The large figure is the Median time to first response. Beside it, the share Answered within a day and how many applications are Still waiting, shown in red when any are.
A first response is the first action your team takes on an application: moving the candidate to another stage, or scheduling an interview. Opening or reading the application does not count.
How long the first reply took sorts applications into buckets from under an hour to over seven days. Applications with no response at all are listed separately as No response yet, rather than as the slowest bucket, because they are still waiting.
Below the figures, Opener answered and the Opener round chart show how candidates responded to the Opener, the AI-drafted first-contact questions. Each round is in exactly one state: Pending, Recorded as sent, Reminded, Answered, Acknowledged, No response or Failed. The labels describe what the product recorded, not what a mail server confirmed was delivered. See Opener: the first contact.
Rejected counts the applications from the range that have been moved to the Rejected stage.
Roles: which jobs need attention?
The figures at the top count Live requisitions (open or paused jobs), Applications in range, live jobs with No applicants yet, and Stalled jobs.
Roles with no recent movement lists open and paused jobs that have applicants but where nobody has moved a candidate to another stage for 14 days or more. It only appears when there is at least one.
The Requisition scoreboard puts every job side by side: status, days open, applications, qualified candidates, average fit, how many reached interview, hires, median time to hire, and how long since anything moved (Idle). It is ordered by the number of applications. Click a job's name to open it.
Exporting these figures
Every chart card has a Chart and Table switch and a CSV download, and every chart here can go into a PDF report. The scoreboard and the stalled list have no CSV button of their own; add them to a report instead. See Reports, exports and schedules.