What is a good LinkedIn connection acceptance rate?

Expect somewhere between 2 and 4 in 10. Across 42 Sellinger customer teams, 39% of invitations were accepted in the 90 days to 19 September 2026. Vendor datasets with millions of requests report 26% to 29%. Definitions and audiences differ, so compare your own rate to your own history, not to one number.

Evidence levels used below: our data* (measured in Sellinger and described with its method), vendor claim (a tool vendor reports it, and has a product to sell), documented (LinkedIn says it). LinkedIn publishes no acceptance benchmark. Everything else here is a report from someone with a dataset.*

What acceptance rate should you expect?

Evidence level: our data. Across 42 customer teams, about 4 in 10 invitations were accepted, 39% in the 90 days to 19 September 2026. We count it per lead, and internal and test teams are excluded. The window is fixed, so it will drift as we re-run the query.

Our own outbound account is a second data point. In the Sellinger case study, 68,818 connection requests went out from up to 43 profiles of real team members, ambassadors and students, and 41% were accepted (Sellinger case study). The audience was founders, heads of growth, sales and revenue leaders, and agency owners. That is one audience, written to in one voice. A campaign to another audience can land well above or well below it.

One more example, to show the spread inside a single customer. A foundry campaign in the Lamothermic case study had 27% of 3,345 requests accepted (Lamothermic case study). Same tool, different audience, different number.

What do third party studies report?

Evidence level: vendor claim. We read the following pages on 19 October 2026. Each defines the rate slightly differently, and none says it is a LinkedIn figure.

They disagree, and the reasons are mundane. The datasets are different audiences. One may count an accept only inside a window, another any accept later, and the pages do not say which. "Good" in a vendor article is also a marketing line. Expandi calls 30% to 45% good and 45% to 60% excellent, bands with no stated dataset.

The AiSDR test is small, and its authors call the gap between 37.3% and 34.6% noise at that volume. We agree. It is useful for what it did not find, which is a large note effect.

What moves acceptance, and what barely does?

Evidence level: mixed, and we say which in the table.

Three points follow from this.

Audience beats copy. The only factor with a big, consistent gap in these sources is who you write to. Overloop's own reading is that below about 30%, targeting or the profile is more likely the problem than the wording. Our view is that they are right.

A note changes replies more than acceptance. Belkins reports 9.36% replies after accept with a note, against 5.44% without. AiSDR reports 19.2% against 2.6%, with a much smaller group. The size of that lift differs by a factor of several, so do not plan on a number. The direction is the same in both. If you want to try no note, our playbook (no note invite, specific opener) shows a sequence that holds the first message back until the person accepts.

People look at your profile before they accept. LinkedIn describes a complete profile as something that will "increase your discoverability and profile search appearances" (LinkedIn Help: build your profile). That is not an acceptance statistic. It is a reason to fix the headline, photo and first lines of the About section before you spend invitations.

How should you read your own number?

Evidence level: our view.

• Separate campaigns by audience. One blended number hides a 40% list and a 15% list.
• Count per person. Divide people who accepted by people invited, once each.
• Wait for the tail. Many invitations are accepted days or weeks after they are sent. A number measured after three days understates the final rate.
• Compare against yourself. A drop from 40% to 25% after you change the list is a signal. A gap between you and a vendor figure is not.
• Watch pending invitations, not only the rate. LinkedIn lists invitations that were "ignored, left pending or marked as spam" among the reasons for a restriction (Types of restrictions). It publishes no threshold. See how the limits work and what the rules say about automation.

A checklist for when the rate falls below your own norm:

• [ ] Did the list change, such as a new industry, a new seniority band or a new country?
• [ ] Did the sender profile change, such as a new photo or headline?
• [ ] Are many invitations still pending after two weeks?
• [ ] Is the reason to connect the same one sentence for everyone?
• [ ] Is the audience people who could plausibly buy, or only people who are easy to find?

What does Sellinger do with this?

Evidence level: our setting. Campaigns can start from intent signals, such as people engaging with a topic, so more of the people invited may already be interested (Intent Signals). The agent writes in the sender's name and voice, and a person can approve every message before it goes. We watch acceptance as well as volume, and we do not publish a restriction rate, because we have not measured one we would stand behind.

What can't we tell you?

• That 39% is your rate. It is an average across customers who chose their own audiences.
• That a note, or no note, will win. The studies point to a small effect either way.
• A safe floor. LinkedIn gives no number below which an account is restricted.
• That a high rate means a good campaign. Acceptance is the first step. Replies and meetings are what pay.

If you want one number to start with, plan for about 3 to 4 in 10 on a well chosen list, and treat anything under 2 in 10 as a prompt to look at the audience first.