Three Touches, Not Ten: What 4 Million Recruiting Messages Reveal About Getting a Reply
When I was laid off from Lumen, I sent a lot of messages into what felt like a void. Cold emails to people I'd met once at a conference. LinkedIn notes to hiring managers I'd never spoken to. A follow-up, then another, then a third that I now recognize was probably annoying. I had no idea which of it was working, because nobody was going to hand a laid-off telecom employee the internal data on what makes a stranger reply to a message.
It turns out that data exists. Nobody collects it by studying job seekers, they collect it by studying recruiters, because recruiters send outreach messages at a scale no individual job seeker ever will, millions of them, tracked, tested, and optimized down to the day of the week. I went and pulled the actual numbers, and the honest version of this piece is: this is recruiter-to-candidate data, not the reverse. But the thing being measured, a stranger deciding in about two seconds whether to respond to an unsolicited message, works the same regardless of which direction it's pointed. If it's what makes recruiters' messages land, it's worth knowing before you send your next one.
The channel gap is bigger than most people assume
Pin, a recruiting outreach platform, analyzed more than 4 million messages sent between June 2025 and May 2026 across 1,500-plus organizations. The channel gap wasn't close. LinkedIn messages pulled a 17.08% reply rate. A recruiter-written email pulled 6.31%. A fully automated email pulled 4.96%. In a cleaner, matched test where the exact same 165,000 people were contacted both ways, 16.6% replied on LinkedIn versus 4.4% on email, essentially the same message, nearly four times the response.
| Channel | Reply rate | Note |
|---|---|---|
| LinkedIn message | 17.08% | Matched-cohort test: 16.6% |
| Recruiter-written email | 6.31% | A human wrote it, not a template |
| Automated email | 4.96% | Sequence-tool default |
| Same 165K people, email arm | 4.4% | Matched-cohort test |
Email volume has also gotten a lot noisier since the platforms optimized for it. Cold email replies overall sat around 8.5% back in 2020; the 2026 baseline is closer to 3.4%, roughly half. Inboxes filled up with automated sequences, and everyone's reply rate paid for it, including recruiters'. LinkedIn hasn't been immune to the same crowding either, but it's still the channel where a message from an actual person is most likely to get seen and answered.
Want someone building your warm-path outreach for you, not guessing at it?
Get 3 free roles →Three touches, not ten
This is the number that actually changed how I coach a stalled outreach plan. Pin's sequence data shows reply rates by touch: 5.52% on the first message, 5.40% on the second, 4.21% on the third, 3.34% on the fourth. Add up where the replies actually come from, and the first three touches capture 93.2% of everything you're ever going to get back. A fourth message only adds another 4.5% on top of that. A separate dataset from the recruiting tool Ashby found a similar shape from a different angle: roughly 7% reply rate after one email, about 15% after two, around 23% after three, and then it flattens out.
If three follow-ups haven't gotten a reply, a fourth, fifth, or sixth almost never will. That effort is better spent on a new person entirely.
The instinct when you're job searching and desperate for a response is to keep pushing the same thread. The data says the opposite: three well-spaced touches, then move on. Persistence past that point doesn't read as diligence to the person on the other end, it reads as someone who isn't picking up the signal.
What actually makes a message worth answering
Format and personalization matter more than most people budget time for. Messages that included the recipient's first name got a 5.13% reply rate versus 2.61% without one, nearly double, from the same 4-million-message dataset. A separate analysis of 31 million emails by the tool Hunter found that going a step further, referencing two specific, personalized details instead of just a name, lifted replies from 3.6% to 5.6%, a 56% increase. On the AI side, fully AI-generated outreach with no human editing pulled 4.1% replies in one 2026 study by the sales tool Saleshandy, while human-written messages pulled 10.4%, and a human-edited AI draft landed at 14.7%, better than either alone.
Length works the same way everywhere it's been measured. LinkedIn's own data shows InMails under 400 characters, a couple of sentences, get about 22% above-average response, while ones over 1,200 characters run 11% below average. Pin's email data found a similar sweet spot around 150-199 words with a five-to-six-word subject line. The instinct to explain your whole situation, every job you've held, why you're a great fit, in the first message is understandable and it's working against you. Short and specific beats thorough almost every time.
Timing, with an honest caveat
Pin's data shows Wednesday and Thursday outperforming other weekdays, and Saturday performing worst, which lines up with ordinary intuition about when people are actually checking messages. It also shows a strange-looking spike for messages delivered between midnight and 5am ET, peaking 3-5am. That's very likely an artifact of when automated sequence tools are scheduled to fire, landing at the top of an inbox before the day's other messages pile in, not proof that 3am is a magic hour. The reliable, low-risk takeaway is simpler: send on a Tuesday through Thursday morning if you have a choice, and don't sweat the exact hour beyond that.
What this means for your next message
None of this is about applying to more roles. It's specifically for the moment you're reaching out to an actual person, a hiring manager, a former colleague, someone one connection removed who could make an introduction. A few things worth changing based on this:
- Default to LinkedIn over email when you have the choice. The gap is roughly four to one in the data above. If you only have someone's email, that's fine, but don't skip a LinkedIn message just because email feels more formal.
- Use their actual name, and one real specific detail. Not "I noticed your background," something true and particular: the actual team, a specific post they wrote, the specific role. It nearly doubles the odds of a reply and it takes thirty extra seconds.
- Keep it short. A few sentences beats a paragraph. Say who you are, why you're reaching out to this person specifically, and one clear, easy-to-answer ask.
- Plan for three touches, then stop. Space them out, days apart, not hours. If you get to three with no response, that's real information, not a reason to send a fourth.
- If you use AI to draft it, edit it yourself before sending. The data on hybrid AI-plus-human outreach beating either pure version on its own held up across every dataset that measured it.
This is also most of what I actually do when I'm building warm-path outreach for a client, the same short, specific, three-touch structure recruiters use on candidates, aimed the other direction. It's not a secret formula. It's mostly just brevity, a real name, and knowing when to stop.
The honest bottom line
This data was built to help recruiters get replies from candidates, not the other way around. I'm not aware of a study that measured job seekers cold-messaging hiring managers at this scale, and I'd be skeptical of anyone who claimed one existed. But a reply is a reply, and the psychology of a stranger deciding whether to respond doesn't flip depending on who's doing the asking. If it's what gets recruiters answered, it's a reasonable bet on what gets you answered too, and it's a lot better than the guessing I was doing at my own kitchen table a year ago.