In short: 25 to 40 seconds, three beats, one question, no pitch. Send it at step two rather than as the first touch. Tested across 5,915 prospects: reply rate 30.7 to 33.7 percent, meeting rate 4.1 to 5.0 percent.
The three beats
Every voice note that worked had the same three parts and nothing else. Say who you are and why you are in their inbox. Say the one specific thing you noticed. Ask one open question and stop talking.
What kills a voice note is the fourth beat, where you explain what your company does. It is the same finding as the text tests: a question-led message with no pitch replied at 44.5 percent against 18.6 percent for the same message with solution detail added. Speaking the pitch out loud makes it worse, not better, because the listener can hear you reading.
Voice against video, and when to use each
| Format | Reply rate | Meeting rate | Sample |
|---|
| Text first touch | 28.8% | 4.1% | 8,385 prospects |
| Personalised video first touch | 40.4% | 6.4% | 8,385 prospects |
| Text follow-up at step two | 30.7% | 4.1% | 5,915 prospects |
| Voice note at step two | 33.7% | 5.0% | 5,915 prospects |
Video is the stronger format and the expensive one. Voice is the cheap one that still moves the number, which is why it belongs at step two on every sequence rather than only on your top accounts.
Frequently asked questions
How long should a LinkedIn voice note be?
Between 25 and 40 seconds, which is roughly 65 to 100 spoken words. LinkedIn allows up to 60 seconds. Anything past about 40 seconds gets abandoned, and anything under 15 sounds like you did not think about it before recording.
Do LinkedIn voice notes actually work?
Modestly, and cheaply. In a test across 5,915 prospects at 10 to 50 employee companies, adding a voice note at step two lifted reply rate from 30.7 to 33.7 percent and meeting rate from 4.1 to 5.0 percent. Personalised video was stronger, at 28.8 to 40.4 percent replies, but costs considerably more to produce at volume.
Should the voice note be the first touch or a follow-up?
Follow-up. We tested it at step two, after a text first touch, and that is where the lift came from. Video is the format that worked as a first touch. A voice note from a stranger with no context is easy to ignore.
Should I script a voice note or improvise?
Script the shape, not the words. Read a written script aloud and it sounds read. Know your first line, the one thing you are asking, and your last line, then talk. The generator gives you those three anchors and timing marks rather than a word-for-word script.
Do voice notes work everywhere?
Not in DACH. Our DACH programmes needed a formal tone and voice notes underperformed there. They worked best with founders and COOs at companies under 50 people, which is where we ran the test.
