Hiring well: structured interviews and scorecards

Many bad hires can be traced back to the interview itself: each candidate got different questions, the interviewer went with a gut feeling and nobody wrote down why one person beat another. Structured interviews and scorecards fix much of that. They make hiring more consistent, easier to defend and, according to the evidence, better at predicting who will do the job well.

This guide covers job analysis, structured questions, scoring rubrics, work samples and the UK legal basics, with a worked example for a small business.

Start with job analysis

Before you write a single question, work out what the job needs. Job analysis means looking at the tasks, the outcomes expected and the skills and behaviours that separate good performance from poor performance. Talk to people who do the job now, or who manage it, and look at what has gone wrong in the role before.

The US Office of Personnel Management (OPM) notes that interview questions based on job competencies identified through job analysis are linked to higher validity and better agreement between interviewers. In UK terms, the output is a job description and a person specification. Acas advises that hiring decisions should be based on how each applicant met each point in those two documents.

From the analysis, pick a short list of competencies to assess. Four to six is a workable number for most roles (a rule of thumb, not a standard). More than that and interviewers start skimming.

Structured versus unstructured interviews

OPM describes interviews as sitting on a scale. At one end, unstructured interviews have no constraints on the questions and end with an overall impression. At the other, every applicant gets the same predefined questions and follow-ups, and answers are scored against benchmarks of proficiency.

OPM reports that interviews with more structure show higher validity, better reliability between raters and less adverse impact on different groups. It also notes that interviewers and applicants tend to prefer less structured formats, which is one reason they persist.

The CIPD’s selection methods factsheet describes a structured interview in similar terms: the same predefined questions in the same order, scored against criteria agreed in advance. It recommends a panel instead of a single interviewer, and says each interviewer should score answers independently before any group discussion. Independent scoring stops a common problem: if the most senior person speaks first, everyone else tends to drift towards their view.

Write the questions

OPM describes two main question types, both built on real situations from the job:

• Behavioural questions ask the candidate to describe something they have done. “Tell me about a time you had to deliver a project with less budget than planned. What did you do?” These rest on the idea that past behaviour predicts future behaviour. OPM notes they have shown higher validity for complex, professional and managerial jobs.
• Situational questions describe a scenario and ask what the candidate would do. “A customer emails to say their order arrived damaged two days before a gift deadline. How do you respond?” These suit candidates with less work history.

Write one or two questions per competency, plus agreed follow-up prompts such as “What was your role specifically?” or “What happened next?” Every candidate gets the same main questions. Follow-ups can vary only within the agreed set.

Build a scorecard with anchors

A scorecard lists each competency, the question or questions that test it, and a rating scale. The scale works best with anchors: short descriptions of what an answer at each level looks like. Without anchors, a “3” means something different to every interviewer.

An example (illustration only). A small online skincare brand is hiring a customer service lead. One competency is “resolving complaints”. The scorecard reads:

Each panel member scores every competency alone, with a one-line note of the evidence, before the panel discusses. The discussion focuses on large gaps between scores. The final decision cites the scores and the evidence, not “she felt like a good fit”.

Weight competencies if some matter more than others, but decide the weights before you see any candidates. Changing the weights after the interviews is how bias gets back in.

Add a work sample

An interview tells you how someone talks about work. A work sample shows you some of the work. The CIPD factsheet describes work sample tasks, simulations and assessment centres, with examples such as a customer role-play, analysing a dataset or drafting a client email. It states that these predict job performance better than tests, education, job history or unstructured interviews.

Keep work samples short, relevant and the same for every candidate, and score them against a rubric. Acas says unpaid work trials should be a reasonable length, usually no more than one day. If the task gives the employer real value, or is not part of the actual recruitment, the candidate may need to be paid the National Minimum Wage. For the skincare brand, a 30-minute task replying to three sample complaint emails is enough.

Stay within the law in the UK

This is a summary, not legal advice.

• Acas states that employers must not discriminate when advertising, interviewing or deciding who to employ.
• GOV.UK lists the protected characteristics under the Equality Act 2010: age, gender reassignment, marriage and civil partnership, pregnancy and maternity, disability, race (including colour, nationality and ethnic or national origin), religion or belief, sex, and sexual orientation.
• Acas says that by law you must ask anyone invited to interview whether they need reasonable adjustments to attend, such as step-free access, extra time for a written test or a sign language interpreter.
• Acas recommends asking each applicant the same questions where possible, having more than one interviewer, and checking interviewers understand discrimination law.
• Interview notes and recordings are personal data. Store them securely and follow data protection rules.

A structured process helps here. If a rejected candidate asks why, you can point to the criteria, the questions and the scores.

Common mistakes, and when to bend the rules

• Writing questions after the interviews are booked. Do the job analysis first.
• “Culture fit” as a criterion. Without a definition it becomes “like us”. Name the behaviours you mean.
• Scoring at the end of the day from memory. Score each answer straight after the interview.
• Discussing candidates before scoring. Independent scores first, then the conversation.
• Dropping structure for internal candidates or referrals. They need the same process, or the comparison is unfair.
• Puzzle questions with no link to the job. They test something, but not the competencies you listed.

Structure can feel stiff, and candidates still need to experience you as human. Leave time at the end for their questions, explain what happens next, and say when they will hear back, as Acas recommends. For very senior or unusual roles, you may add a less structured conversation about direction and values. Keep it separate from the scored interview and do not let it override the evidence.