Reply rate is the most useful single number in outbound recruiting, because it fails loudly. Bad targeting shows up as silence within days, long before it shows up as an unfilled role. It is also the number most distorted by vendor marketing — case studies quoting 60% reply rates are describing warm audiences, tiny samples, or a different definition of "reply".
What follows are benchmark ranges from practitioner reports across the teams and tools we track, the levers that move the number in the order they actually matter, and the measurement traps that make dashboards lie. Treat the ranges as directional: they vary by market, seniority, and season, and they are best used to spot outliers in your own data.
Benchmarks by channel
| Channel | Typical reply rate | Notes |
|---|---|---|
| Personalised email, tight list | 15–35% | The realistic target for outbound done properly |
| Templated email, broad list | 3–8% | Volume play; costs brand and deliverability |
| LinkedIn InMail | 10–25% | High attention, credit-limited; declines count as replies |
| LinkedIn connection request + note | 20–40% accept, far lower conversation rate | Accepts are not replies — measure conversations |
| Referral or warm intro | 50–80% | Highest-converting channel by a wide margin, and not scalable |
| SMS / WhatsApp | 25–45% | High response, high irritation risk; check consent rules in your jurisdiction |
| Phone (cold) | 5–15% connect | Still effective for frontline and trades roles |
Benchmarks by seniority and scarcity
| Audience | Expected reply rate | Why |
|---|---|---|
| Early career (0–3 years) | 25–40% | More open to opportunity, less outreach saturation |
| Mid-level, common skills | 20–30% | The healthy middle where most hiring happens |
| Mid-level, scarce skills (ML, SRE, security) | 8–18% | Heavily saturated inboxes |
| Senior / staff+ engineering | 8–15% | High saturation, high bar for relevance |
| Director and above | 10–20% | Lower volume of outreach, but only from credible senders |
| Frontline / non-desk | 15–30% by phone or SMS | Email is often the wrong channel entirely |
The pattern that matters: saturation, not seniority, drives the number down. A staff-level ML engineer receives more recruiter contact in a week than a director of finance does in a quarter. Judge your performance against the saturation of the specific population, not against a global average — and against your own trend, which is the only truly comparable series you have.
The five levers, in order of impact
1. Targeting (largest by far). Contacting 30 well-chosen people beats 300 plausible ones on replies, interviews, and brand. The mechanism is simple: relevance is visible in the first sentence, and irrelevance is too. Most "our messaging isn't working" problems are list problems. The list construction we recommend is in the passive candidate playbook.
2. Evidence of real research. One concrete, specific reference to their actual work — a project, a talk, a migration, a product they shipped. Generic flattery ("your impressive background") measurably underperforms because it signals a mail merge. This is the lever automation most often gets wrong; when evaluating agents and sequencers, read what they actually generate.
3. Compensation transparency. Including a range raises replies in most markets and filters mismatches before they consume interview slots. Pay-transparency legislation is making it expected anyway; vagueness now reads as a low offer.
4. Follow-up count. Going from one touch to three roughly doubles total replies in practitioner reports — the cheapest available improvement, and the most commonly skipped. Cap at three; four-plus unanswered touches buys resentment.
5. Message craft (smallest). Length, structure, subject line. Worth getting to "short, specific, one small ask" and then leaving alone. Teams that A/B test subject lines while sending to a loose list are optimising the last 5% of a number the first lever controls.
Timing, volume, and the tradeoff nobody states
Send-time optimisation is oversold: mid-week mornings in the recipient's timezone are marginally better, and the effect is small relative to targeting. The volume tradeoff is much more important and is rarely made explicit.
Reply rate and send volume move in opposite directions, because scaling sends means loosening targeting. A team going from 50 to 400 messages a month will usually see reply rate fall from the 20s into single digits. Whether that's a good trade depends on capacity: more total replies is only progress if someone answers them within hours. Interest from a passive candidate decays fast, and unanswered positive replies are the most expensive leak in outbound sourcing.
There is also a cost that doesn't appear on any dashboard. Every low-relevance message spends a little of your employer brand in the exact market you'll recruit from for the next five years — and bounced sends spend your domain reputation, quietly lowering deliverability for the outreach that is well-targeted.
Measuring it honestly
Four traps make reported reply rates useless:
Counting only positive replies. Count every human response including declines and "not now". Excluding declines inflates the number and hides the fact that people are reading it.
Mixing warm and cold. Referrals and rediscovered past applicants reply at multiples of cold rates. Report them separately or your cold benchmark is fiction — and see the rediscovery guide for why warm audiences deserve their own motion.
Ignoring bounces in the denominator. A 4% bounce rate means your real reply rate on delivered mail is higher than reported, and your deliverability problem is worse than you think. Track both. The contact-data test in our accuracy guide is the fix.
Attributing at the campaign level only. Reply rate per role and per recruiter is where the actionable variance lives. One saturated role can drag a whole team's average and hide the fact that everything else is working.
Reply rate is one of the five numbers we'd manage the function with; the others are in our metrics guide. And remember reply rate is a means, not the end: a 40% reply rate on a badly targeted list produces conversations that go nowhere, which is why sourced-to-interview pass-through belongs next to it on the same report.
What tooling changes and what it doesn't
Engagement platforms — Gem, SourceWhale, Interseller — reliably improve two of the five levers: follow-up discipline (nothing gets dropped) and measurement (per-role, per-recruiter reporting you can act on). Both are real gains, and neither fixes targeting.
AI and agentic tools can improve the research lever if — and only if — the personalisation is grounded in actual profile evidence rather than templated praise. That is the specific thing to inspect in a pilot: pull twenty messages the tool actually sent and read them as a candidate would. Tools whose outreach you'd be embarrassed to sign are a brand liability regardless of their reply-rate claims.
No tool fixes a loose list, an uncompetitive offer, or a role nobody wants. Reply rate is honest about all three, which is exactly why it's worth reporting. For the tool field itself, see the 2026 ranking.