How to Read a CV for Evidence, Not Vibes

You've got fifteen CVs open in fifteen tabs and forty minutes before your next meeting. You skim each one for maybe twenty seconds, feel a vague pull toward three of them, and move on. Two weeks later you can't remember why you liked any of them.

That's not reading a CV. That's forming an impression and calling it a decision. A CV is a claims document. Your job is to read it the way a bank reads a loan application — not to be cruel about it, but because vague claims and specific claims are not the same thing, and only one of them tells you anything.

Separate claims from evidence

Every line on a CV is one of two things: a claim about identity or a claim about what happened.

• "Results-driven marketing professional" — identity claim. Worth nothing. Nobody writes "mediocre marketing professional."
• "Grew organic traffic from 8,000 to 34,000 monthly visits over 11 months, working with one freelance writer" — evidence claim. This tells you the starting point, the outcome, the timeframe, and the resources involved.

The first kind is filler. The second kind is checkable and specific enough to ask about. When you're reading a CV, mentally grey out every adjective and bold every number, date, and named outcome. What's left is what you're actually evaluating.

Most CVs are 80% adjective and 20% evidence. That ratio is itself information — it tells you how much digging the interview will need to do.

The four things a bullet point should tell you

A strong bullet answers, even implicitly:

1. What was the situation? (team size, budget, market, constraints)
2. What did this person specifically do? (not "the team achieved" — what did they do)
3. What changed as a result? (a number, a shipped thing, a resolved problem)
4. Over what time? (three weeks and three years are very different achievements)

When a bullet is missing two or more of these, don't assume the worst — assume it's unfinished, and hold the question for the interview. "Led a project that improved customer retention" is missing the scale, the mechanism, and the timeframe. That's not a lie. It might just be a rushed CV. But it's a gap, and gaps are where your interview questions should go.

Watch for verbs that hide the actor

"Managed," "oversaw," "contributed to," "involved in," "participated in" — these verbs are doing a job, and the job is ambiguity. They let someone describe being in the room without saying what they did in it. Compare:

• "Involved in the migration to a new CRM" — could mean they led it, could mean they attended two meetings about it.
• "Configured the CRM's lead-scoring rules and trained six sales reps to use them" — unambiguous.

When you see a hiding verb, that's your interview question, verbatim: "You wrote you were involved in the CRM migration — what part of that did you personally do?" Don't rewrite it to sound less pointed. The plainness is the point.

Read the shape, not just the lines

Once you've read the bullets, step back and look at the shape of the whole document:

• Does responsibility grow over time, or does it plateau? Someone who's had the same scope of work for six years across three job titles is telling you something different from someone whose scope keeps expanding.
• Do the dates on the CV match the story the bullets tell? A "led a 20-person rollout" claim sitting inside an 8-month role at a 15-person company is worth a gentle question, not an accusation.
• Is there a pattern in what's left out? No numbers anywhere, ever, across every job — even in roles where numbers are normal (sales, ops, support) — tells you either the person doesn't think in outcomes or didn't have access to them. Both are useful to know before you hire them into a role that requires exactly that.

None of this is about catching someone in a lie. Most CVs aren't dishonest — they're just written in the genre of CV-speak, which rewards confident vagueness over precise detail. Your job is to translate it back into facts.

Build a two-column read

A method that works without any special tooling: split a page in half. Left column, copy the claim exactly as written. Right column, write either the evidence that supports it or a one-line question you'd need answered to believe it. Do this for the four or five strongest claims on the CV, not all of them — you're not auditing, you're triaging.

By the time you finish, you'll have a short, specific list of things to ask in the interview, and it'll be a much better list than "tell me about yourself." You're not trying to trip anyone up. You're trying to turn a document full of assertions into a conversation full of specifics.

What a CV can't tell you, no matter how well you read it

Be honest with yourself about the ceiling here. A CV, even read carefully, tells you what someone says they did — not what they can do next, under your constraints, with your team. That's why a well-read CV should shrink your interview time, not replace it, and why a work sample or a structured skills check earns its place for anything the job actually depends on. Some hiring tools cap how much a CV can weigh in the final decision for exactly this reason — a document written by the candidate, about the candidate, will always flatter the candidate a little. Reading it well is about getting the most honest signal out of it before you move on to evidence it can't provide.

A CV read for evidence takes longer than a CV read for vibes. Twenty minutes instead of twenty seconds, for the shortlist that matters. But you'll walk into the interview with real questions instead of small talk, and that's the difference between an interview that confirms a hunch and one that actually tests it.