What 500k LinkedIn requests say about booking meetings

Hey friends, every few weeks a 45% acceptance rate screenshot does the rounds on LinkedIn, with a lot of fire in the comments.

Nobody asked how many of those people replied.

So we went and looked at what an acceptance usually turns into. The answer is mostly silence.

In today’s Circuit:

🛠️ Operator Tip of the Week
📖 What 500k LinkedIn connection requests say about meetings
💡 Tool Spotlight
📡 Across the Wire
🧪 Worth trying this week

Pull your acceptance rate by segment this week and pause anything sitting under 20%. SmartReach found about 1 in 4 campaigns landed below that line, and together those campaigns ate roughly 30% of all invitations sent.

LinkedIn gives each rep around 100 invites a week, so a segment that won't accept is burning your scarcest resource. Usually the fix is upstream: wrong titles, wrong company size, or a list that was never a fit.

SmartReach just published its State of LinkedIn Outreach report, built on its own campaign data from January to June 2026: 1,800+ campaigns, 480+ reps, and more than 500,000 connection requests.

It's one of the cleaner looks we've seen at what happens after someone clicks accept, and that part of the funnel is where most of the leak is.

The average acceptance rate across those campaigns was 26%, with the top quartile clearing 40%. That's the number people screenshot. Follow it down the funnel though, and it shrinks fast.

Out of 1,000 invites, about 260 accept, about 47 reply, about 12 reply with any real interest, and 3 to 4 book a meeting. Put another way, 80% of accepted connections never send a single reply, and 11% of campaigns that got acceptances produced zero replies at all.

Accepting a request costs the other person one tap, but replying means they have to want something from you.

There's a long-running fight about whether to send connection requests blank or with a note, and the data gives both camps something. Blank requests got accepted more often (28% versus 24%).

Once accepted though, people who'd received a note replied at 22% versus 14% for blank, about 1.5x. Run both through 1,000 invites and the note comes out ahead: roughly 53 replies against 39, if those two rates hold together on the same audience (the report gives them separately, so treat that as our back-of-envelope math).

The likely reason is that a note gives the conversation somewhere to start. But free LinkedIn accounts only get 3 to 5 personalized notes a month, so this lever mostly belongs to people on Premium or Sales Navigator.

Follow-ups behave in a way that surprises most people. SmartReach measured the lift over a single message: the first follow-up added roughly nothing, the second added about 4 percentage points, and from the third onward the gains faded to around 1 point.

Spacing the touches 3 to 5 days apart worked best. So a lot of teams are stopping exactly one message too early, after the follow-up that barely moves anything and right before the one that does. Stacking six more after that mostly annoys people (who would have thought?).

This is probably no surprise, but the biggest gap in the whole report sits in the channel mix. Per 1,000 prospects, LinkedIn alone booked about 4 meetings. Adding email took it to about 14, and adding calls on top got to about 21.

That's roughly 5x the meetings from the same prospects. Reply rates showed a 1.6x lift going multichannel, and the uplift varied a lot by industry: about +36% in recruiting, +53% in SaaS, and +90% in e-commerce and retail.

Some of that gap probably comes from the kind of teams that run three channels (more process, better lists), so treat it as a strong signal with some noise in it. Still, if LinkedIn is your only channel, it's the first thing we'd look at.

AI-written messages. Across the campaigns, AI-assisted messages replied about 0.3 percentage points differently from manual ones, which is basically no difference, and around 40% of campaigns were already using AI.

The tool you write with matters a lot less than who you're sending to, whether there's a note, and whether you show up anywhere besides LinkedIn.

At least that’s what the data says. We believe that obvious AI slop will eventually perform worse.

HeyReach is a LinkedIn automation tool built around running many sender accounts from one place, with a shared inbox for all the replies.

With each rep capped at around 100 invites a week, HeyReach lets a team spread a campaign across several senders, and it plugs into email tools, which is how you get to the LinkedIn-plus-email setup that booked more than 3x the meetings in this report.

But more senders means more invites going to the same list, so it multiplies whatever targeting and messaging you already have, good or bad, and running lots of accounts through automation always carries some risk of LinkedIn restrictions.

- Expandi's 2026 benchmark (13.2M connection requests) found acceptance held steady at 28-30% all year, but the reply rate on connection notes slid from 3.5% to 2.2%, a 37% relative drop. People still accept, they just answer the note less. Read here

- HeyReach looked at 96,051 campaigns and found the ones running 6 to 20 senders had the strongest reply performance. Scaling volume dented acceptance slightly, but replies held up. Read here

- Belkins' follow-up study (7.5M emails, 15M LinkedIn touches) found the first email has the best per-step reply rate at 0.59%, but follow-ups together bring in 58.6% of all replies. Read here

- Split one segment in half for two weeks, blank requests against a short note, and then compare replies per 100 invites.

- Add a second follow-up 3 to 5 days after the first if your sequence currently stops at one.

- Add an email step for anyone who accepted but never replied.

- Pause any segment running under 20% acceptance and look at the list before you look at the copy.