How to Interview an SDR for Cold Email Skills in 2026
By Puzzle Inbox Team · May 22, 2026 · 10 min read read
The exact questions, role-plays, and scoring rubric to interview SDRs for cold email skills in 2026. Hire the operator, not the resume.
What "good" looks like when you interview an SDR for cold email in 2026
The job has changed. In 2026, an SDR who only knows how to "send sequences in Outreach" is a junior. The bar for a hireable cold email SDR now includes deliverability instincts, list-building judgment, copy taste, and reply-handling craft. This guide gives you the exact interview loop we use to hire SDRs who hit quota in month two.
If you are going to interview an SDR for cold email skills, stop asking "tell me about a time you handled rejection." Ask what their bounce rate was last month and why.
The four competencies that actually matter
- List quality judgment - can they spot a bad ICP filter?
- Copy taste - can they tell why an email will bomb before sending?
- Deliverability instincts - do they know what kills a domain?
- Reply craft - can they turn a "not interested" into a meeting?
Round 1: the 30-minute screen
Five questions, no fluff. Each answer should take 3-5 minutes.
Question 1: "Walk me through your last campaign, end to end."
What you want to hear: ICP definition, source (Apollo, Clay, scraped), enrichment waterfall, verification, segmentation logic, send volume, inboxes used, warmup status, copy variants, and reply rate. If they cannot name their tools and numbers, end the call.
Question 2: "What was your bounce rate last month and what did you do about it?"
Good answer: under 2%, with a verification step named. Great answer: they mention catch-all handling and role-account filtering. Red flag: "I do not track that."
Question 3: "Show me a subject line you wrote that worked and one that did not."
You are testing taste, not template recall. The best SDRs talk about why the subject line worked - specificity, curiosity, lowercase, no clickbait.
Question 4: "How do you decide between a dedicated sending tool and a sequencer like Apollo?"
See our Apollo sequences vs dedicated piece for the framework they should articulate. They should know the answer depends on volume, deliverability SLA, and team size.
Question 5: "What does a healthy warmup look like?"
They should mention gradual ramp, SPF/DKIM/DMARC, separate domains for sending, and the rule of thumb of 30-50 sends per inbox per day at steady state.
Round 2: the take-home (90 minutes, paid)
Give them a real prospect list of 25 contacts (you can pull from Hunter or AnyMailFinder) and ask for:
- A two-sentence ICP critique - is this list any good?
- A three-step sequence written for the segment
- Three subject line variants per step
- A one-paragraph note on what they would A/B test first
Pay them $150. Candidates who refuse a paid take-home are not candidates - they are tire-kickers.
Scoring rubric
| Competency | Weight | What to look for |
|---|---|---|
| ICP critique | 20% | Spots at least one filter problem |
| Copy specificity | 30% | Names a trigger event, not "saw your website" |
| Subject lines | 20% | Lowercase, under 6 words, no salesy verbs |
| A/B test logic | 15% | Picks one variable, calls out sample size |
| Tone | 15% | Sounds like a human, not a template |
Round 3: the live role-play
This is where you separate operators from resume-writers. Two role-plays, 15 minutes each.
Role-play 1: the reply triage
Hand them ten realistic inbound replies (interested, not-interested, wrong-person, auto-reply, hostile, "send info," "what is the price," forward-to-colleague, out-of-office, and a vague maybe). Ask them to triage in real time and draft responses to three. You are watching for tone, speed, and judgment.
Role-play 2: the "not interested" save
You play a prospect who replies "not interested." They have one email to earn a meeting. This is the single highest-signal exercise we run. Top performers respond with curiosity and a specific, low-friction ask. Weak candidates either grovel or push.
What to test for AI/automation fluency in 2026
Every SDR you hire in 2026 should be comfortable with AI sequence tooling. Ask: "How do you use Regie.ai or similar tools without sounding like a robot?" The answer should include human-in-the-loop editing, never "I let it write and send."
The references question that works
When you call references, ask one question: "What were their reply rates compared to the team average?" If the reference cannot answer or says "I do not know," you have learned everything you need to know.
Where Puzzle Inbox comes up in interviews
If a candidate has run multi-mailbox campaigns and mentions a unified inbox tool like Puzzle Inbox for reply management, that is a green flag - it means they have actually scaled past two inboxes.