Free blueprint, PDF, 7 pages
Message patterns that get replies, and the ones that don't
We checked popular LinkedIn message advice against 29,385 real messages, comparing each sender only with themselves. Three patterns held. Six popular rules did not, and five fresh templates show what to do instead.
Free. We'll email you the PDF. Questions about the data: hello@zen-mode.io.
What's inside
- Why the sender outweighs the copy (acceptance ran from 9% to 67% across 18 seats), and how we controlled for it
- What held: a note on the invite (+7.4 points on replies, same sender), and no fade by follow-up 3
- What didn't: note length, ending on a question, extra merge fields, Tuesday to Thursday, questions and links in follow-ups
- Five templates written fresh for this guide, each tied to a pattern that held
How it was put together
29,385 messages from 79 LinkedIn seats on ZenMode, 8 April to 23 September 2026, with internal and test accounts removed. Every pattern is tested within the same sender with a 95% interval; cells with fewer than 100 messages or 10 outcomes are left out. Aggregates only, and the templates are fictional.
The longer write-up is on the blog: LinkedIn Connection Request Messages That Actually Work (With Examples).
ZenMode runs your LinkedIn outreach from your own computer, on your own internet connection. Your computer has to be on while campaigns run. Start your 14-day free trial