Interview Skills

Why interviewers say your answers were too vague

Being told your interview answers were too vague is almost never a vocabulary problem. It means you described the category of work you do instead of one thing that happened. Here is how to hear the difference in your own answers and repair them before the next round.

AM

Andre Mendes

·

August 24, 2026

·

12 min read

The note from the recruiter said the panel wanted more specific examples, and you read it twice, because you thought you had been specific. You talked about de-escalating families at the bedside, about covering two units the week the float pool came up short, about the survey that came back with no findings. It felt concrete while you were saying it, which is why the note reads as though somebody wrote it about a different interview.

When an interviewer says your answers were too vague, they rarely mean your words were unclear. They mean you described a category of work instead of one thing that happened. Replace each summary with a single event with a date, a place, another person in it, and something that was different at the end.

What interviewers actually mean by "too vague"

Nobody is confused by "I coordinate patient flow" or "I manage stakeholder expectations." Those sentences are perfectly clear. They are also unusable, because the interviewer is trying to build a short film of you doing the work, and a category of activity cannot be filmed. Only an occasion can.

That is why "be more specific" is such useless advice. You already believe you were specific, and by your own standard you were: you named the unit, the function, the type of problem. The mismatch is altitude. You answer at the height of a job description, and the interviewer needs you down at the height of a Tuesday.

STAR gets recommended at exactly this point, and it survives the criticism fine, since plenty of candidates who can recite all four letters still answer vaguely. The trouble sits in two of them. Situation gets filled with a standing condition, "we were chronically short staffed," where it wants a single day. Result gets filled with a quality, "it went well," where it wants something that was measurably different afterward. Get those two loaded correctly and the framework does the rest of its job on its own.

It matters because of what happens after you leave. If somebody has to describe you to a colleague who was not there, a story survives that retelling and a summary does not, and when a hiring manager can repeat nothing you said, the safest thing available to them is that you were fine.

The four shapes vagueness takes

Vagueness arrives in four recognizable shapes, and each one repairs differently. They also keep company with four other answer habits worth testing yourself against.

The answer that describes the role instead of an event

This one usually lands on the first behavioral question, when nerves push you toward the safest thing you know how to say, which is your own job description.

"In my current role I'm responsible for coordinating patient flow across the unit, keeping staffing ratios safe, and escalating to the physician team when something changes."

Everything there is true and none of it is evidence. The repair is one shift.

"On a Thursday night in February we were three nurses short on a twenty-eight bed medical floor. I called the house supervisor before eight, had two floats within the hour, and told the hospitalist which patients I expected to call about overnight. We finished with no missed medication passes."

The answer that merges five occasions into one

The composite answer is sneakier, because it sounds like a story. It cannot carry details, though, since you are describing an average of several real events rather than any one of them. The tell is the word "whenever." Picture a high school English teacher, call her Priya, asked how she handles a parent who disputes a grade.

"Whenever a parent challenges a grade, I hear them out, I walk them through the rubric, and usually we get to a place where they understand the reasoning."

"In October a father emailed the principal about his daughter's essay score and copied me, saying the rubric was subjective. I asked for fifteen minutes before school, brought her essay along with two anonymized essays from the same assignment, and had him grade all three himself before I showed him my marks. He landed close to me on each, then asked what she could do to move up."

The move Priya made there is a good one, and it is also beside the point. The requirement is that the occasion can be checked, and cleverness has nothing to do with whether it can. Here is the same October with none of the ingenuity in it.

"In October a father emailed the principal about his daughter's essay score and copied me, saying the rubric was subjective. I asked him to come in before school, sat with the rubric and her essay in front of both of us, and showed him where the argument lost points. He still thought two of the criteria were soft. He also agreed the score was the score, and he asked me what she could do to move up."

The answer built out of adjectives

Quality words feel like content because they are the words job postings and performance reviews use, which is why they surface under pressure.

"I'd say I'm collaborative and very detail oriented. I work well cross functionally and people know they can rely on me to catch things."

An adjective is a conclusion, and conclusions are the interviewer's to draw, so hand over the thing the conclusion came from and let them do it.

"Our monthly close kept slipping because the three field offices coded travel expenses differently. I sat with Dana in accounts payable, mapped every code her team saw in March against our chart of accounts, and found the whole travel category landing in one catch-all bucket. I rewrote the submission template, and April close finished two days earlier than March."

Nobody in that answer says "detail oriented," and everybody who hears it thinks it.

The answer that hedges with "typically" and "would"

Conditional grammar is a reliable sign that no single real instance is behind the answer. "Typically," "usually," "I would," and "we'd" all answer a question about the past in the tense of a policy.

"Typically, if a deadline was going to slip, I would flag it early with my director and we would usually re-sequence the work."

"In March our permit backlog hit six weeks and the council packet was due at the end of the month. I went to my director as soon as I knew we were going to miss it, with two options: pull both plan reviewers off intake for a week, or push the commercial applications to April. She picked the reviewers, we cleared the packet, and I called the applicants who got pushed and gave each of them a date."

Sometimes the hedge is protective, because you have not faced that exact situation. That is a different problem with a different fix, and it is worth reading how to answer when you genuinely lack the experience before you reach for it.

Four questions to test your own answers

Record an answer on your phone and play it back against these four, which is more reliable than reading your notes, because reading hides the hedges.

• Could a stranger draw this? If someone with no context could sketch the room or the hallway, you told a story. If they could only draw a flowchart, you told a job description.
• Is there a date and a place? A month and a location force your memory to load one real event rather than an average of many.
• Does anyone in it have a name or a role? No house supervisor, no father in the front office, no Dana in accounts payable, and it is a process you are describing.
• Did something change by the end? It does not have to be a win. A number moved, a decision got made, or you carried something into the next one.

An answer that fails all four is a different genre of speech, and rewording never turns it into evidence.

Priya's plainer version passes all four exactly as cleanly as her clever one: you can draw the two of them at a desk before first period, it happened in October in a specific school, the father and the daughter are both in it, and by the end a man who had gone over her head was asking her what to do next.

How do you remember details from years ago?

Most of what you think you have forgotten is written down somewhere you already have access to. Open your own calendar for the month it happened, and the meeting titles will hand you the date, the room, and who else was in it. Search your sent mail for the thread, because the version of events you wrote at the time is more reliable than the version you have been rehearsing since. Search Slack or Teams for the person's name and you will find the hour it escalated and the sentence you actually used. Your last performance review already contains the numbers somebody else thought were worth writing down about you, which is a useful check on the ones you would have guessed. The status report or the handover note you wrote that week is the densest source of all, since it was written for the sole purpose of telling somebody what had changed. And the shift schedule or the project plan will tell you how many people were on and how long you had, which is the scope figure most people otherwise invent. Thirty minutes of that produces real numbers for three stories, which is roughly what one round asks of you.

How many facts one answer actually needs

Three, all of them checkable: one time anchor, one number that shows the scale of the thing, and one outcome. The other person comes free, since occasions have people in them. That is a cap and not a target, and it exists because precision buys credibility only up to the point where it starts sounding rehearsed. Six dates in a row reads as a script, and once a panel decides you memorized the answer, they stop listening for what is inside it.

Approximation clears the bar. "Just after shift change" locates an event as usefully as 7:20 does, and "sometime that spring" and "about forty percent" are both specific enough to build an answer on. The only test that matters is whether a follow-up question would find something real underneath, which is exactly where approximating and inventing part company.

The most common coaching note by a wide margin

Across 7,582 improvement notes written on more than 3,700 individually scored answers (2,712 scored answer blocks inside live sessions, 550 standalone practice answers, and 439 anonymous practice answers, queried 2026-08-25), the largest single theme is a request for more specific evidence, at about 22 percent. The second largest, at about 9 percent, is a request to quantify the result. Themes were grouped from the note titles the coaching model wrote, so the line between those two is softer than the percentages make it look, and roughly three in ten notes are some version of "make this concrete."

The literal titles repeat almost word for word: "provide specific examples," "add specific examples," "quantify the outcome." The people receiving them come from a paying base that skews toward nursing and healthcare, education, government and public safety, management, and finance, so this is not a habit that belongs to one field. There is a full breakdown of why strong candidates give weak interview answers if you want the rest of what the scoring found.

What to do when the outcome genuinely cannot be quantified

"Quantify the outcome" is fair advice in sales and unfair advice almost everywhere else. If you are a hospice nurse or a public defender, the result of your best work is a person, and no honest percentage attaches to it. Inventing one is worse than having none, because it is the one part of the answer a follow-up question can take apart.

Quantify the scope instead. Scope is factual, and it does the job the metric was supposed to do, which is to show the interviewer the size of the room you are used to standing in. Picture a probation officer, call him Ray, who opens with the fact that he carries fifty-five active cases and then tells you about the three who violated in the same month: that gives a panel more calibration than any satisfaction score would, and every word of it is true.

Then say what changed in the terms the work actually uses. Ray has no outcome percentage and does not need one, because he has the man who kept every appointment for six months after missing three in a row, and a panel can hear the distance between that and "I build strong rapport with my clients."

Frequently asked questions

How specific should an interview answer be?

Specific enough that a follow-up question has somewhere to land, and short enough that the detail never becomes the answer. Ninety seconds to two minutes is the working range for a single event. If it runs long, cut the backstory that got you to the moment before you cut anything that happened inside it, since the backstory is the part nobody can check and the part nobody remembers.

How many specific stories do I need for one interview?

Three or four, recovered properly, will cover most of a round, because one occasion answers several questions depending on which part of it you lead with. The night you were three nurses short is a prioritization story or a communication story, and which one it becomes is decided by your first sentence. Build the three from detail you can actually recover, then choose at the table.

What if the only feedback I got was "be more specific"?

Treat it as a note about structure. Pull the answers you gave that day and check each one for whether it named an occasion or described a category. Assume the phrase is a compression of something longer, and that the thing underneath it was an inability to picture you doing the job.

What if I cannot name the client or the details are confidential?

Say "a regional health system in the Midwest" or "a mid-size credit union" and keep everything else intact. Confidentiality removes proper nouns, not the scene. What does get noticed is when confidentiality becomes the reason an entire answer stays at ten thousand feet.

Listen for the word "typically"

Agreeing with this changes nothing while you are sitting there, because vagueness is a speaking habit that returns the moment a stranger is looking at you, and "In my current role I'm responsible for..." arrives on its own. The only way to know whether your repaired answers hold is to say them to somebody who asks the follow-up question when a detail is missing.

If you have a round coming up, run a practice interview on the job you are actually interviewing for and then count how many times you said "typically," "usually," or "I would." Each one marks a place where the answer climbed back up to the height of a job description while you were talking, and you can hear it on a recording even when you could not feel it in the room.

AM

About the author

Andre Mendes — Founder, Mock Interview Pro

André is the founder of Mock Interview Pro, where job seekers run realistic mock interviews and get coach-level feedback on every answer.