Most lists of “good interview questions” are written for the candidate. Search for questions to ask an interviewee and the results are candidate-prep advice: what a strong answer sounds like, how to phrase your own response under pressure. None of it tells a hiring manager what to actually put in front of a candidate, in what order, or which of those questions the law doesn’t let you ask in the first place.
This is the version for the other side of the table.
Why our role guides stopped reusing questions
When we started writing hiring guides for the roles our customers hire most, from Account Executive to Warehouse Manager, the interview-questions section was the one we got wrong first. Early drafts pulled from the same pool of behavioural prompts and swapped the job title: a generic “tell me about a conflict with a colleague,” reused across roles with nothing in common. It read fine and measured nothing. That question doesn’t tell you whether a Sous Chef can run a pass under pressure, and it doesn’t tell you whether an Electrician will stop work when the wiring doesn’t match the documentation.
So we rewrote every guide’s question set from the role’s own failure modes instead, and tagged each prompt to the specific trait it is meant to surface. For a Customer Support Representative, one prompt we kept is: “Tell me about a situation where a customer or guest was openly angry. How did you react, and how did the interaction end?”, scored against empathy and tonality. For an Electrician, the comparable prompt is: “Describe a situation where you noticed a safety or standards violation on a construction site or during a customer job. What did you do?”, scored against safety and standards awareness. Different jobs, different failure modes, different questions.
Each of our 51 role guides now carries a question bank built around what that specific job gets wrong.
Build the set before the first interview
Start from the role’s own traits
Decide what you are testing before you write a single question. Three to five traits is usually enough for an SMB hire: the technical core of the job, one or two behavioural traits that predict how someone handles the role’s actual pressure, and whatever tends to go wrong when the wrong person gets hired into it. The method behind this is the structured interview: job-related questions, a set order, and the same rating criteria applied to everyone. The evidence for why that outperforms a free-flowing conversation is in our longer piece on the research; the short version is that the questions matter more than who is asking them.
Mix behavioural, situational and role-specific questions
A behavioural question asks for a real past example: what happened, what the candidate did, how it turned out. A situational question poses a scenario specific to the job (the till is down and a queue is forming, a deploy just broke production at 2 a.m.) and listens for how the candidate reasons through it live, in real time, rather than how well they recite a textbook answer. For technical or trade roles, a short work-sample task beats another interview question outright: watching someone trace a wiring fault or debug a failing endpoint tells you more in twenty minutes than an hour of conversation does. Each of the role guides linked below ships a mix of these, tagged to the trait it tests, so you are not guessing which type fits which job.
Score it before the debrief
Scorecards are where the question set turns into a decision instead of a vibe. Set the questions and the rating criteria for the role once, then every interviewer records their own assessment independently, before anyone compares notes out loud. In Join, a scorecard stays private to its author until published; the candidate’s profile then shows each published assessment and, once more than one has been submitted, the average rating per attribute. The software does not decide who gets hired. It gives every interviewer the same form, so a debrief compares evidence from one shared event rather than four separate interviews that happened to cover the same candidate.
Know what you are not allowed to ask
Which questions are off-limits depends on where you are hiring, and the two most common lists for Join’s customers are not the same one.
In the UK, the Equality Act 2010 makes it unlawful to discriminate in the arrangements an employer makes for deciding who to offer a job to, against any of nine protected characteristics: age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race, religion or belief, sex, and sexual orientation (Equality Act 2010, s.4). Acas guidance is sharper than the statute on the point that trips up the most interviewers: do not ask about health or disability at any stage of recruitment, interview included, before a conditional offer has been made, outside narrow exceptions for reasonable adjustments, positive action or a genuine occupational requirement.
In Germany, the Allgemeines Gleichbehandlungsgesetz (AGG) protects a shorter list of seven grounds: race or ethnic origin, sex, religion or belief, disability, age, and sexual identity (§1 AGG). §7 AGG bars disadvantaging a candidate on any of them, including on the interviewer’s mere assumption that one applies. Marital status and pregnancy are not separately named grounds under the AGG the way they are under the UK Act; both still reach a German employer indirectly, through sex discrimination, but the two statutes genuinely do not list the same things, which is exactly why “we followed best practice” is not a substitute for checking the law of the country where the role sits.
We sell interview software here, so weigh this section accordingly: it is a starting map, and a real compliance review is a separate, qualified job. Anything touching health, family status or immigration in a question set is worth a conversation with someone qualified in that jurisdiction before you finalise it.
What we’d tell a first-time interviewer
Keep the set short: three to five questions per competency is plenty, where a dozen just tests the same thing from a different angle. Decide what you will not ask before the candidate sits down, so the legal screen happens in prep rather than mid-sentence. And start from the role itself; a job-specific question set already exists for most of the roles an SMB hires.
Each of the 51 guides below pairs behavioural, situational and technical prompts with what a strong and a weak answer actually sounds like, scored against the trait it is testing.
Sales. Account Executive, Head of Sales, Sales Development Representative, Sales Engineer, Sales Manager.
Software and data. Backend Engineer, DevOps Engineer, Frontend Engineer, Full-Stack Engineer, Data Analyst.
Marketing. Growth Marketer, Marketing Manager, SEO Manager.
Product, project and design. Operations Manager, Product Manager, Project Manager, Product Designer, UX Researcher, Content Manager.
People and HR. Head of People, HR Generalist, Recruiter.
Finance and admin. Accountant, Financial Controller, Payroll Specialist, Administrative Assistant, Executive Assistant, Office Manager.
Customer support and implementation. Call Centre Agent, Customer Success Manager, Customer Support Representative, Implementation Manager.
Engineering and manufacturing. Electrical Engineer, Mechanical Engineer, Project Engineer, Production Supervisor, Quality Manager, Warehouse Manager.
Logistics and transportation. Dispatcher, Fleet Manager, Logistics Manager.
Hospitality and healthcare. Hotel Front Desk Manager, Restaurant Manager, Sous Chef, Dental Assistant, Medical Receptionist, Practice Manager.
Facilities, trades and education. Caretaker, Cleaner, Electrician, Teacher.
