LinkedIn connection requests in 2026: what the data says, and 8 templates that fit it

Almost everyone has an opinion about LinkedIn connection requests, and most of them are wrong. Should you add a note or not? How long should it be? Does personalization really move the needle? The arguments go in circles, but the data from 2025 and 2026, drawn from studies of tens of thousands of tracked requests, is far clearer than the debate. This is what actually gets accepted now, and how to turn it into requests that not only connect but start conversations.

Start with the benchmark, because you can't judge your outreach without one. Across B2B outbound, the average connection request acceptance rate sits around 30 to 37 percent. Above 40 percent is genuinely good, and elite campaigns with tight targeting reach 50 to 60 percent. The number that matters most, though, is the floor: a sustained acceptance rate below 20 percent isn't just underperformance, it's a signal to LinkedIn that your targeting is off, and it's one of the fastest ways to get your account restricted. If you're under 20 percent, stop and fix your list before you send another request.

Now the finding that surprises most people. Across multiple 2026 datasets, requests sent with no note at all consistently accept higher, often 55 to 68 percent, than requests with a generic note, which land anywhere from 15 to 45 percent. A blank request is low-friction and reads as a simple 'let's connect.' A generic note ("Hi, I'd love to connect and share what we do") reads as the opening line of a pitch, and people decline pitches. So if your note is generic, you are actively hurting yourself. You would be better off sending nothing.

But acceptance was never the note's real job. The note exists to start the conversation, and here the data flips: adding a message nearly doubles the post-acceptance reply rate, from around 5.4 percent with no note to 9.4 percent with one, in one large dataset. And a genuinely personalized note holds acceptance at 45 to 60 percent, close to a blank request, while setting up the reply. So the real rule isn't 'note versus no note.' It's this: a generic note is the worst of both worlds, a blank request is a safe default, and a specific personalized note is the best overall outcome, because it wins the accept and the reply.

That raises the obvious question: what counts as personalization? The data is blunt about it. Using someone's first name gives no meaningful uplift over a generic message, both sit in the 15 to 25 percent range. What works is specificity anchored to something real: referencing a recent funding round, a role change, or a specific post the person published lifts acceptance into the 50 to 60 percent range. Actually engaging with their content before you send, a thoughtful comment on a recent post, pushes acceptance past 60 percent, versus roughly 30 to 35 percent for first-name-only on the same audience. Specificity is the entire lever. Everything else is noise.

Length matters more than people expect, and shorter wins. The character limit is 300 on every LinkedIn plan, but the best-performing notes run 120 to 180 characters. The principle is simple: the longer the note, the more it reads like an automated pitch. Say one specific, true thing and stop. A two-line note that references their actual work beats a paragraph that lists your credentials every single time.

Here's the lever most people ignore entirely: who you target moves the number more than what you write. The same message gets wildly different results by seniority. Individual contributors accept 60 to 72 percent of requests, managers 50 to 62 percent, directors and VPs 38 to 50 percent, and C-level executives just 25 to 38 percent. Industry matters too, recruitment and HR clear 40 percent, tech and software sit around 35 percent, finance and banking closer to 25 percent. If your acceptance rate is low, the problem is usually your list, not your copy.

Timing is a smaller but free edge. Requests sent Tuesday through Thursday, between roughly 8 and 11 AM in the recipient's local time, outperform, and business hours beat evenings by more than 20 percent. None of this matters if the targeting and message are wrong, but once those are right, sending at the right time is a free few points of acceptance.

One more piece ties it all back to account health. Everything above only compounds if your account stays alive. A standard account can safely send roughly 80 to 100 well-targeted requests a week without tripping restrictions, and staying above that 20 percent acceptance floor is both a performance goal and a safety one. If you want the full picture, see the real LinkedIn limits in 2026 and how account restrictions actually work.

With the data in hand, here are eight templates, organized by the situation you're actually in. None of them are meant to be pasted verbatim, they're structures. The whole point of the data above is that the specific detail is what works, so treat the brackets as the real content, not filler.

1. Engaged with their content (the highest performer, 60 percent plus). "Your post on [specific topic] stuck with me, especially the point about [specific detail]. Been following your thinking on this, glad to connect." This works because you've referenced a real, specific thing, and ideally you commented on that post first.

2. Shared group or event. "We're both in [group / at event] and your take on [specific topic] stood out to me. Would like to stay connected beyond it." Genuine shared context is one of the few cases where a note reliably beats a blank request.

3. Trigger-based (funding, role change, launch). "Congrats on [the raise / the new role / the launch], the [specific relevant angle] caught my eye. Keen to follow what you build next." Trigger-based personalization is what pushes acceptance into the 50 to 60 percent tier.

4. Mutual connection. "[Mutual name] and I were just talking about [topic], and your name came up as someone deep in it. Would value connecting." Real, verifiable, and it borrows trust without pitching.

5. Peer or community. "Fellow [role] here. Your point on [specific topic] matched something I've been wrestling with. Connecting to keep learning from your posts." Peer framing lands especially well with individual contributors, your highest-acceptance segment.

6. No note (the honest default). When you genuinely have nothing specific to say, send the request with no note. The data is clear that a blank request beats a generic one, so don't manufacture filler. Silence outperforms a pitch.

7. Post-decline soft re-approach. "No worries if the timing's off, I mainly wanted to connect because of your work on [specific thing]. Either way, enjoying your posts." Low pressure, references the real reason, leaves the door open.

8. Re-engaging a dormant connection. "It's been a while, [name]. Saw your [recent post / news] and it reminded me we'd lost touch. Would be good to reconnect." Specific and warm beats a generic 'just checking in.'

Notice the thread running through every one of these: they only work if the bracketed detail is real and specific, and that is exactly what breaks at scale. Nobody can genuinely research and personalize 80 to 100 requests a week by hand, so most people fall back on the generic note the data says to avoid. That gap is the whole reason to put an AI agent on it. With Crispy, your agent reads each prospect's recent posts and profile, writes a note that references something true, or sends none when there's nothing worth saying, and does it inside safe daily limits. Templates give you a floor; an agent personalizing each one at scale is how you reach the 60 percent tier without burning your week. See how the AI SDR workflow runs, explore prospecting, or start with Crispy.