Short answer: across 389,890 prospects and 15 controlled tests, a text-only first touch replied at 28.8 percent and a personalised video first touch at 40.4 percent, with 57 percent more meetings. Targeting mattered more than either: broad role-based targeting replied at 13.0 percent against 51.9 percent for a verified ICP and subject-matter match. Most teams benchmark against the wrong number and conclude outbound is dead. It is not.
Ask five sales leaders what a good LinkedIn reply rate is and you get five different answers, all defended with confidence and none with data. Having managed outbound teams for years before building Prospectio, I have been guilty of it myself. So here is what the numbers behind the product actually show.
Reply rates by message type
| Variant tested | Reply rate | Meeting rate |
|---|---|---|
| Text-only first touch | 28.8% | 4.1% |
| Personalised video first touch | 40.4% | 6.4% |
| Broad role-based targeting | 13.0% | 1.4% |
| Verified ICP plus subject-matter match | 51.9% | 6.5% |
| Company-led copy | 10.8% | 1.0% |
| Question-led copy | 33.2% | 4.5% |
The gap between the first row and the last is the entire story of modern LinkedIn outbound. It is not a tuning difference, it is a different sport. The mechanics of why voice and video pull these numbers are in our piece on voice notes at scale.
The numbers in front of the reply rate
Reply rate gets the attention, but two numbers upstream decide most of the outcome.
Accept rate first. If fewer than about a quarter of your connection requests are accepted, the problem is targeting, not messaging. Strong campaigns sit well above that because they qualify before they send: right role, right company profile, right moment. This is why Prospectio scans every profile against your ICP before a word goes out; skipping poor-fit prospects raises every number downstream and protects account health at the same time.
Then follow-up depth. A large share of replies arrive after the first message, which surprises nobody who has done the job and is ignored by almost everyone running a tool. Sequences that stop after one touch leave most of their meetings on the table. The campaigns that perform run multi-step sequences, warm-up actions, a voice note after the accept, message, video, email, with sensible delays between steps.
From reply to meeting
A reply is not a meeting. The leak between “interested” and “booked” is the most expensive one in outbound, because the hard work is already done. The reply sits in an inbox overnight, the prospect cools off, the thread dies. Speed to response is the fix: reply while the interest is live, handle the obvious questions, propose a time. It is exactly the part a human is worst at, checking an inbox at 11pm, and the reason we built an auto-responder that books the meeting end to end.
How to use these benchmarks
Run your own funnel against the chain: accept rate above a quarter, reply rate that clears the template band, replies answered within the hour, meetings booked from a third or more of interested replies. Wherever your number falls furthest below the line, that is the constraint. Fix that one first; fixing anything else just feeds a leak.
And treat any tool’s headline numbers, ours included, as a range you verify on your own pipeline, not a promise. The honest version of benchmarking is that your ICP, offer and market move every figure. The pattern that holds everywhere: qualify before sending, sound like a person, follow up more than feels natural, and answer fast.