Comparing interview scores with performance after 90 days

You can set up a structured interview properly, award scores, and still find yourself thinking afterwards: “Why did that score of 4 for problem-solving skills turn out to be wrong?” The moment your selection process really improves is when you link interview data to performance in the workplace. The 90-day review is a practical starting point for this: close enough to keep memories fresh, yet long enough to really see someone at work.

What do we mean by ‘90-day feedback’?

The 90-day feedback is a brief, standardised review of an employee’s performance after approximately three months in the role. It is not part of a formal appraisal cycle, but rather a checkpoint that can be used to gather insights for your recruitment process.

In blue-collar roles, engineering, logistics and sales, three months is often exactly the point at which you can see whether someone has got the basics under control: safety, pace, learning curve, teamwork and reliability in keeping to agreements.

Important: you should not use this feedback to “prove yourself right” about a hire, but to calibrate your selection tools. This makes comparing interview scores with actual performance a concrete and repeatable process.

Why comparing interview scores with performance works

If you never look back, an interview score remains nothing more than an opinion with a number attached to it. By linking performance after 90 days to the scores from your interview, you can see which factors are predictive and where there is noise.

This helps in three ways:

• Level of difficulty: Which questions actually yielded evidence?
• Competence level: Which criteria correlate with performance in this role?
• Process level: In what areas do interviewers consistently score differently (calibration problem)?

It dispels two common misconceptions

• “We got on well, so it’s all right.” A ‘click’ can co-exist with mediocre performance.
• “That candidate was just a bit of a fluke.” Sometimes, yes, but you often spot patterns when you compare several hires.

Step 1: Decide which performance indicators you will measure after 90 days

Start small and focus on specific roles. It’s better to choose 6 clear indicators than 20 vague points. Distinguish between output, behaviour and learnability.

Examples by job type

• Production / operator: error rate, safety alerts, speed relative to the standard, availability on routes, call quality.
• Technical / maintenance: first-time-fix, fault analysis, systematic working practices, documentation, safety and lock-out/tag-out discipline.
• Logistics: picking errors, speed, adherence to work instructions, reliability, cooperation with planning.
• Sales: activity metrics (calls/appointments), pipeline quality, follow-up discipline, early-stage conversion, CRM hygiene.

Make indicators observable

For each indicator, write one sentence: “What does a 3 look like, and what does a 5 look like?” This ties in directly with the concept of rubrics in a selection scorecard.

Are you already using a scorecard? If so, take your existing structure and keep it consistent. The blog Create a selection scorecard for hiring managers helps to formulate criteria in a clear and measurable way.

Step 2: Decide which interview scores you are going to compare

Only use scores that you have already recorded consistently during the interview. If you start coming up with new labels afterwards, it will skew your data.

Practical set of interview prompts

• Competency scores (e.g. 1–5) from your scorecard
• Case score (separate, so you can compare it with “real-life problem-solving”)
• Disqualification criteria (yes/no): driving licence, certification, shift work, etc.
• Notes for each question: brief facts, no interpretations

If you don’t yet have a fixed interview structure, it’s difficult to make a fair comparison later on. The approach set out in How to conduct a structured interview in 8 steps provides you with the basics for making interview data usable.

Step 3: Collect 90-day feedback without any hassle (but in a professional manner)

The quickest way is to use a short form to be completed by the line manager, supplemented by a brief self-reflection from the employee. Keep it to 10 minutes per person.

Example: 90-day questionnaire (manager)

• Which tasks can this employee carry out independently? (select from the list)
• Rate on a scale of 1–5: quality, pace, safety, teamwork, ease of learning
• What was the biggest risk you anticipated, and did it materialise?
• What needs to improve over the next 30 days to reach the “good” level?

Privacy and retention periods

You process personal data here and, in some cases, performance review information. Ensure transparency: let staff know that you use feedback both for their development and to improve the recruitment process.

Rules regarding retention periods and purpose limitation also apply to application data and assessments contained in the recruitment file. The Explanation from the Dutch Data Protection Authority regarding job applications helps to organise this neatly.

Step 4: Simplify the equation using a small table

You don’t need to be a data scientist. Start by creating a table where you compare the interview scores for each new hire with their performance over the first 90 days. You’ll often quickly spot where your assessment criteria are too optimistic or where a question adds little value.

This table is designed to help you quickly spot patterns for each criterion.

What is “good enough” to draw conclusions?

With just one or two hires, you’re mainly seeing random results. With 6–10 hires in the same role, you’ll usually start to see useful patterns. From 15–20 onwards, you can start to look seriously at patterns by interviewer or by question.

Research shows that structured interviews are, on average, better at predicting outcomes than unstructured interviews, but the quality depends heavily on your scoring and the questions you ask. An accessible summary of selection validity can be found at the The American Psychological Association on structured interviews and selection.

Step 5: Turn insights into improvements in your interview

The benefits almost always lie in small adjustments. Asking just one more question can make more of a difference than adding a whole new skill.

Three common issues (with specific solutions)

• Your interview overemphasises ‘presentation’. Add evidence-based questions: “What steps did you take, and what was the measurable result?” and award fewer points if the answer is just empty rhetoric.
• Your case study does predict outcomes, but interview questions do not. Give the case study greater weighting on your scorecard, or create two shorter versions of it.
• Interviewers vary considerably in their scoring. Plan a short calibration session and re-evaluate what a 3 versus a 5 means, as described when working with a structured interview.

Link this to your candidate journey

If you notice that new recruits with lower “trainability” are more likely to drop out or take longer to settle in, don’t just look at the interview questions. Also check your onboarding and mentoring processes. A good selection process without a good start remains vulnerable.

Common mistakes in the 90-day feedback process

• Do not use fixed definitions. “Good” means something different to every team leader. Address this using behavioural anchors.
• Using feedback as an assessment tool. That’ll just lead to defensive responses and less honest feedback.
• Picked up too late. After six months, memories become selective and roles have often changed.
• Do not translate back into ‘process’. In that case, it remains an HR ritual with no impact on the quality of recruitment.

Practical next step

If you want to compare interview scores with performance, start with a single role and a single quarter’s worth of hires. Set up a brief 90-day feedback process, create a single summary table, and spend 30 minutes each month discussing what you’ve learnt.

If you want to implement this data-driven approach without it becoming a separate project, FosFor can help you refine scorecards, interview rubrics and a measurable 90-day evaluation tailored to blue-collar, sales and maritime roles. Take a look at our approach at These 4 mistakes are costing you your best candidates and work out where you can make the most profit.