"LinkedIn reply rate: why the same 13 replies are 8% or 25%"
NudgeLink team · Published · Last updated · 3 min read
Short answer: there is no single LinkedIn reply rate. The number depends on what you divide by, and the same results can honestly be called 8% or 25%. Pick one denominator, write it next to the number, and when someone quotes you a rate, ask what's underneath it.
One month, two denominators
These are the numbers from a client's first month with NudgeLink: 161 connection requests sent, 51 first messages sent after people accepted, 13 replies, 8 of them positive.
| Replies divided by | Calculation | Rate | What it answers |
|---|---|---|---|
| Invites sent | 13 / 161 | 8% | How many conversations does an invite buy? |
| People messaged | 13 / 51 | 25% | How well does the message work on people who accepted? |
Positive replies work the same way: 8 of 161 invites is 5%, and 8 of 51 messages is 16%.
None of these is wrong. They answer different questions, and the mistake is putting one next to a number that answers a different one. A "25% reply rate" in one tool's marketing and an "8% reply rate" in your dashboard can describe identical campaigns.
Acceptance has the same problem. In that month, 47 people accepted. Out of the 141 invites checked so far that is 33%. Out of all 161 sent it is 29%, because 20 invites hadn't been checked yet. We use the first, because an invite we haven't checked isn't a "no", it's an "unknown".
Why the denominator is the thing to watch
Two traps come up again and again.
The flattering denominator. Screenshots and sales pages quote whichever version looks best, usually replies per message sent to people who already accepted. It's a legitimate number. It's just not the one that tells you what a campaign costs.
The denominator that moves. If a rate is divided by something that changes for reasons unrelated to replies, the rate moves with it. The classic case is dividing replies by the people currently in the sequence: people who replied leave the sequence, so the number drops as soon as you add new leads, even though nothing about your replies changed. If your rate changes and you can't point to a reason in the messages, check the bottom of the fraction first.
Count people, not messages
Someone who replies three times is one reply. Someone you message twice is one person messaged. Counting messages makes a reply rate look lower or higher depending on how many follow-ups you send, and it makes campaigns with different sequences impossible to compare.
That's how NudgeLink's dashboard counts: replies divided by the people we've reached with a message or follow-up. LinkedIn messages and emails are kept as separate totals, because they behave differently, but the rate is about people. All replies count, positive, neutral and negative. Positive replies are shown separately.
A related rule: if there's nothing to divide by, the dashboard shows a dash, not 0%. An account with two invites out today doesn't have a 0% acceptance rate. It has no data yet.
The three numbers worth tracking
Write the denominator next to each one.
- Acceptance rate: accepted ÷ invites checked. It tells you about targeting and account health. If it falls, look at who you're inviting first (lead fit matters more than message copy).
- Reply rate: people who replied ÷ people messaged. It tells you about the message. If it falls while acceptance holds, the problem is in the writing (don't pitch in the first message).
- Positive replies per 100 invites. It tells you what the whole campaign buys. Use it for planning: how many invites do you need for the conversations you want, and what pace is safe (how many invites per day).
Read them together. If only acceptance moves, it's targeting. If only the reply rate moves, it's the message. The third one moves with both, and with volume.
Four questions to ask about any reply rate
When a vendor, an agency or a competitor quotes you a number:
- What is it divided by? Invites, accepted connections or messages?
- Are replies counted as people or messages?
- Does it include replies that said no? A "reply rate" of 30% that is half "please stop" is a different thing from 30% interested.
- How recent are the messages? Someone messaged yesterday hasn't had time to reply. A rate that includes this week's messages is lower than the real one.
If they can't answer the first one, the number isn't worth much.