Here's a question worth sitting with before your next sales interview: who is selling whom?
You think you're assessing a candidate. You're asking about their pipeline, their process, their proudest deal. But the person across the table is doing exactly what you're paying them to do to your prospects — reading the room, building rapport fast, shaping a story that lands. If they're any good at the job, they're better at that in the room with you than anyone else you'll speak to all week. And you're about to make a hiring decision based on how well they did it.
That's the part nobody says out loud in sales hiring. It isn't that interviews are a flawed tool clumsily applied. It's that the exact skill set you're hiring for — persuasion, rapport, narrative control, reading an audience and adjusting on the fly — is the same skill set a candidate uses to beat your interview. You are not testing whether they can sell. You are being sold to, by someone with every incentive to perform well and every skill required to do it. The better the candidate would be at the job, the better equipped they are to make your process tell you the wrong thing.
Think about the last sales hire you were genuinely excited about before they started. Chances are the excitement came from the interview itself — the way they handled a tough question, the story about the deal they saved in the eleventh hour, the easy confidence when you pushed back. None of that is evidence they can sell. It's evidence they can interview. Those look identical from across a desk, and telling them apart is exactly the part of the process you skipped.
Confidence isn't a signal. It never was.
Most hiring managers believe they can tell a strong candidate from a weak one by how sure of themselves the person seems. Confident answers, no hesitation, direct eye contact — it reads as competence, so it gets scored as competence.
New research says that instinct is worse than useless. A 2026 reanalysis by researchers at the University of Bath and the London School of Economics, using corrected statistical methods across large replication datasets, overturned the popular version of the Dunning-Kruger effect. The comforting story — that unskilled people are blissfully overconfident while skilled people know their limits — doesn't hold up. The researchers found overconfidence is close to a universal trait, and if anything, the most capable people showed the most of it. Confidence tracks almost nothing about actual ability, in either direction.
Sit with what that means for your interview scorecard. If you're rating candidates on how sure they sound, you're not measuring skill. You're measuring a personality trait that shows up in strong performers and weak ones in roughly equal supply. The quietly spoken candidate who hedges their answers might be the more self-aware operator in the room. You'll likely never know, because your interview was built to reward the other one.
The story is the product, and you already bought it
Ask a sales candidate to walk you through their best deal and watch what happens. You'll get a narrative arc — tension, a clever discovery question, an objection handled with grace, a close. It's polished because it's been told before, in other interviews, possibly to other people entirely.
A 2021 survey of 1,250 American job applicants, run by ResumeBuilder.com, found 32% admit to lying on their resume outright. The most common lies aren't small ones: 46% inflated years of experience, 40% inflated skills, 43% stretched the length of time they held a role. Eighty per cent of the people who lied got hired anyway. The lie only surfaced later — 41% had an offer pulled once it was caught, 18% were fired after starting. Most weren't caught at all.
Now put that number next to a separate, much larger body of research: a 2006 meta-analysis by Charles Bond and Bella DePaulo, pooling 206 separate studies and more than 24,000 individual judges, on how accurately people detect when they're being lied to in real time. The answer: 54% accuracy. Barely better than a coin flip. Humans are not equipped to spot deception in conversation, and decades of research say training doesn't fix it much either.
Put those two findings together and look at what you're actually doing in a sales interview. You're listening to a story from someone with a documented, near-one-in-three chance of embellishing it, using a skill — real-time judgment of truthfulness — that the evidence says you don't reliably have. You're not verifying the story. You're enjoying it. Those are not the same activity, and only one of them belongs in a hiring decision.
Knowing the words isn't knowing the job
Every sales candidate now arrives fluent in the language of the discipline. They'll say MEDDIC, Challenger, consultative selling, "diagnose before you prescribe." It sounds like expertise, because it uses the vocabulary of expertise. It often isn't.
There's a version of this pattern showing up in unrelated fields, and it's worth borrowing because it makes the mechanism obvious. Recent academic research tracking financial analysts found a measurable, negative relationship between how optimistic and polished someone's public self-presentation was and how accurate their actual forecasts turned out to be — and the analysts with the weakest accuracy but the strongest self-presentation were the ones most likely to get promoted. The researchers didn't chalk this up to overconfidence. They called it what it is: professionals learning what the market rewards saying, and saying it, independent of whether it's true.
Sales candidates learn the same lesson faster than almost anyone, because the entire job is built around knowing what a buyer wants to hear. If reciting "I run a MEDDIC-qualified process" gets nods in an interview, candidates will say it whether or not they've ever built a compelling event with a real economic buyer under pressure. You didn't test their process. You tested their ability to describe a process convincingly to someone who wasn't going to ask a follow-up sharp enough to expose the gap.
You never tested the one thing that matters
Here's the honest audit question: in your last five sales hires, how many times did you actually watch a candidate handle a real objection, live, with something on the line? Not describe how they'd handle one. Handle one.
Most sales interviews never get there. They stay in the past tense — tell me about a time, walk me through a deal — which means every answer is pre-written, rehearsed, and impossible to disprove in the room. The one moment that would tell you something real — a live discovery call against your actual product, your actual pricing objection, your actual "we already have a vendor" — gets skipped, because it's uncomfortable to run and slower to build than a list of behavioural questions pulled from a template.
That's the gap this whole article has been circling. The interview measures whether someone can talk fluently about selling, under conditions with zero real risk and total narrative control. The job measures whether someone can sell, under conditions with real risk and a buyer who isn't cooperating. Those are different skills, tested by different formats, and only one of them tells you what you need to know before you sign the offer letter.
What actually works
Stop asking candidates to narrate the past. Put them in the present tense instead. Hand them a real objection from your business — the one your reps hear every week — and make them respond to it live, on the spot, with you playing a difficult buyer. Don't tell them it's coming. The rehearsed answer disappears the moment the scenario isn't the one they prepared for, and what's left is closer to the truth than anything a story can give you.
Score it against specifics, not impressions: did they ask a real diagnostic question before pitching, did they handle the objection or just push past it, did they adjust when you pushed back a second time. Write that scorecard before the interview, not after, so your own impression of how likeable the candidate was doesn't quietly become the score.
Verify what they tell you rather than admiring how well they tell it. A number they can't back up with a name, a manager, or a CRM screenshot is a story, not a credential. Call the reference yourself, and ask about the deal they told you about specifically — not "were they a good employee," which nobody answers honestly, but "what happened on this exact account." Most stories survive a polite reference check. Fewer survive a specific one.
And don't let one senior, confident voice in the hiring process settle the decision alone — the research on overconfidence applies to interviewers too, and the most certain person in the room is not automatically the most accurate one. Score independently, compare afterward, and treat a split decision as information, not an inconvenience to be talked out of.
None of this makes the interview useless. It makes it honest about what it can and can't tell you. A conversation will always tell you whether you enjoyed forty minutes with someone. It was never going to tell you whether they can sell — and the sooner you stop asking it to, the sooner your hiring stops rewarding your best-recruited candidates instead of your best salespeople.
If your interview process is still testing how well someone talks about the job instead of how well they'd do it, that's worth fixing before your next hire — not after.
If you want help diagnosing the real issue — or hiring the right profile for the motion you actually run — talk to SalesHQ, or download the FY2027 Sales Salary Guide.
Ready to hire or take the next step in your sales career?