Creating assessments
An assessment is built in three layers: the questions themselves, the sections that group them, and the settings that decide how the whole thing behaves.
Sections
Sections group related questions — "SQL", "JavaScript", "Problem solving". Candidates see them as separate parts, and results are reported per section, which is what makes them worth using even when you only have a handful of questions. A section-level breakdown tells you what a candidate is weak at; a single overall score does not.
Timing
Duration is a whole-assessment limit in minutes, not a per-question one. The clock starts when the candidate begins and does not stop if they close the tab.
A candidate who loses their connection can return to the same link and continue — their answers up to that point are already saved. The deadline is fixed when they start, though, so time spent away still counts against them. If the deadline passes while they are gone, that attempt is over.
You can also set a start date and end date on the assessment. Outside that window the links do not open, regardless of what individual invitations say.
Pass mark
Pass mark is a percentage. It labels a result for your team and appears in exports and in outbound integrations as a threshold value.
It is never applied to the candidate. Candidates are not told whether they passed, no email announces an outcome, and nothing in the platform rejects or filters anyone based on their score. That is a deliberate design constraint, not an oversight — see Automated decisions.
Shuffling
Two independent switches:
- Shuffle questions — randomises order within each section
- Shuffle sections — randomises the order of the sections themselves
Both reduce the value of a shared answer key between candidates sitting the same assessment. Neither is a substitute for a large enough question pool.
Multiple-choice scoring
Three settings control how multiple choice is marked:
| Setting | Effect |
|---|---|
| Scoring method | Partial credit awards marks per correct option selected. All or nothing requires the exact correct set. |
| Negative marking | On by default. Wrong selections subtract marks. |
| Penalty per wrong answer | How much each wrong selection costs. Default is 1.0. |
Negative marking exists to make guessing unprofitable on multi-select questions. If your questions are single-answer, turn it off — it does nothing useful there and candidates find it punitive.
What you ask candidates for
Candidate fields control the intake form shown before the assessment starts, and required fields control which of those are mandatory.
Collect as little as you can defend. Everything here is personal data about someone who is not yet your employee, it is subject to the retention window on your account, and a candidate reading the privacy notice before your assessment sees exactly what you asked for.
The invitation email
Invitation email text replaces the default body of the email candidates receive. Notify users picks which of your team members get an email when an attempt is submitted.
Preview before you invite
Every assessment has a Preview action. It runs the real candidate experience — including code execution against your test cases — without creating an attempt or consuming your plan's allowance.
Preview at least one code question before inviting anyone. It is the fastest way to catch a broken test case, and a broken test case scores a competent candidate zero.