Diagnosing Why Your Interview Did Not Land

Ask most unsuccessful candidates what went wrong and you get a feeling, not a finding. I was nervous. I think I rambled. Station four was a disaster. Those are impressions, and impressions are a poor basis for spending a year fixing something.

Diagnosis is a separate skill from performance and almost nobody practises it. What follows is a method. It takes about two hours, it is unpleasant in the way that useful things often are, and it produces a specific answer rather than a mood.

Step one: rebuild the tape

You cannot diagnose what you cannot see. Sit with a blank page and reconstruct the day in order: every station or question, the prompt as best you remember it, roughly what you said, and what the assessor did while you said it. Write it as description, not judgement. This part is not the review.

Do this within a day or two if you possibly can. Interview memory decays fast and it decays selectively, keeping the two moments that embarrassed you and dropping the six that were simply flat. If weeks have passed, write what remains and accept the gaps rather than inventing detail to fill them.

One detail worth capturing: what the assessor asked as a follow up. Follow ups are diagnostic. A probe that asks you to consider another side usually means you gave one side. A probe that asks for an example usually means you spoke in generalities. A probe that asks what you would actually do usually means you analysed and never decided.

Step two: sort each station into a failure mode

Go station by station and put each one into a category. Most people expect their failures to be scattered. They are usually not. The pattern is the finding.

• Wrong question answered. You gave a strong response to a nearby question. Classic in ethics stations, where candidates deliver a prepared position on consent when the prompt was about resource allocation.
• One sided. You had a view and defended it for four minutes without acknowledging that anyone reasonable could think otherwise.
• No landing. You explored every angle beautifully and never said what you would do. Assessors read this as avoidance, not nuance.
• Thin material. The personal stations had nothing specific in them. Adjectives about yourself instead of events.
• Shapeless. Everything was there and it arrived in the order it occurred to you. The best point turned up after the timer.
• Rehearsed. Fluent, polished, and visibly not responsive to the person in front of you. This scores worse than a rougher answer that is clearly being thought.
• Presence. Fast, quiet, no eye contact, hands doing something strange. The content never got through the delivery.

If five of your eight stations land in one category, you have your answer and it is a single fixable thing. If they scatter evenly, read on, because that means something different.

Step three: mark yourself the way they did

Your categories tell you what happened. The rubric tells you whether it cost you. Understanding how MMI scoring works in Australia matters here because the marking is usually global and domain based rather than a checklist of points, which means a station can feel ragged and still score in the middle band, while a station that felt smooth can score low because it showed no reasoning at all.

Give each of your reconstructed stations a rating out of five on the things that are typically rated: understanding of the issue, quality of reasoning, consideration of other perspectives, communication, and insight. Be strict. The purpose of this exercise is not to feel better.

Then look at the shape of your ratings. A row of threes is the most common and the most misread result. Nothing was bad. Nothing was memorable. In a field where the offer line sits well above average, competent and forgettable is a losing profile, and it is a completely different problem from having one catastrophic station.

Step four: check whether the format was working against you

Some candidates are consistently better in one format than the other. If you sat both a circuit and a panel this year, compare them honestly, and read up on how MMI and panel interviews differ in Australia. People who need a run up struggle in eight minute stations and do better in a conversation that builds. People who talk themselves into trouble over twenty minutes do better in short, contained stations. Formats vary by school and change between cycles, so check the university's current admissions page before you build a plan around one of them.

It is also worth asking whether you understood what you were walking into. A surprising number of unsuccessful interviews come down to a candidate who prepared for a conversation and got a circuit. If any part of what an MMI actually is surprised you on the day, that alone explains a lot, and it is the cheapest thing on this list to fix.

Step five: test your diagnosis against evidence

Self diagnosis is unreliable in a specific way: you can hear your intentions, and assessors can only hear your words. You know you meant to weigh both sides. Whether you said anything that a stranger would recognise as weighing both sides is a separate question, and you cannot answer it from memory.

So test it. Take three prompts of the type you think you handled worst and answer them cold, out loud, on a timer, then have them marked by something that does not know what you meant. MasterMed's live AI interviewer runs timed MMI stations and marks against a rubric, which is a quick way to check whether your suspected weakness shows up when someone else is scoring. The first speaking station is free on the trial, no card, and the trial never converts by itself. If your hypothesis was wrong, better to find out in March than next November.

What not to conclude

Three conclusions get reached constantly and almost none of them survive contact with evidence.

• That the assessor did not like you. Assessors have a rubric and a stack of candidates. Facial expression during a station is not a score.
• That one bad station sank you. Circuits exist precisely to dilute single stations. One weak station among eight rarely decides an outcome.
• That you are not the sort of person they want. This is a story about identity, not a diagnosis, and it prevents the work that would actually change the result.

And the finding you have to allow for

Sometimes the honest answer is that you performed reasonably and a strong field and a limited number of places did the rest. If your reconstruction shows solid stations, sensible reasoning, and nothing you would meaningfully change, the diagnosis is variance rather than deficiency. That points to a different plan: more applications, more interview chances, and a modest sharpening rather than a rebuild.

The point of the whole exercise is to leave with one sentence you can act on. Not I need to do better. Something like: my ethical stations never reached a decision, or my personal answers had no specific events in them, or I speak far too fast for the first ninety seconds. One sentence, testable, fixable. That is a diagnosis. Everything before that is just a feeling with a long memory.