In May 2025, a connection request with a short personal note pulled a reply 3.5 percent of the time. By April 2026, the same move pulled 2.2 percent. Nothing about the message changed. The acceptance rate on those requests did not move either, holding between 28 and 30 percent for eleven of those twelve months. Only the replies dried up.
The number getting quoted out of this year’s benchmark data is a different one: LinkedIn messages pull replies at roughly twice the rate of cold email. That is true, and sales leaders are already moving budget on it. The 3.5 to 2.2 line is the one that tells you how long the advantage lasts.
Key statistics at a glance
| Metric | Value | Source |
|---|---|---|
| Connection requests analyzed (May 2025 to April 2026) | 13,218,869 | Expandi 2026 benchmarks |
| Outbound messages analyzed | 6,730,447 | Expandi 2026 benchmarks |
| Platform-wide connection acceptance rate | 28.5% | Expandi 2026 benchmarks |
| Message reply rate (after acceptance) | 10.4% | Expandi 2026 benchmarks |
| Connection-note reply rate | 3.0% | Expandi 2026 benchmarks |
| Connection-note reply decline over 12 months | 3.5% to 2.2% (37% relative) | Expandi 2026 benchmarks |
| Cold email response rate | 5.1%, down from 7% | Expandi H1 2026 report |
| Reply rate across 15.1M touchpoints | 7.2% | Belkins 2026 study |
| Reply rate with a personalized note | 8.2% vs 5.3% without | Belkins 2026 study |
| Reply rate, AI-assisted messaging | 7.3% vs 7.0% without | Belkins 2026 study |
| Monthly request volume growth | 1.11M to 1.25M (about 13%) | Expandi 2026 benchmarks |
What is a good LinkedIn reply rate in 2026?
Around 10 percent on messages, and just under 30 percent acceptance on connection requests. Expandi’s 2026 benchmark analysis covers 13,218,869 connection requests and 6,730,447 outbound messages sent from 13,302 accounts between May 2025 and April 2026. Platform-wide, that data shows 28.5 percent connection acceptance, a 10.4 percent message reply rate, and a 3.0 percent reply rate on the notes attached to connection requests.
A separate 2026 study from Belkins puts the overall reply rate at 7.2 percent across 15.1 million outreach touchpoints, with an industry range from 4.2 percent to 10.5 percent.
Those two figures are not the same measurement and should not be averaged into one. Expandi’s 10.4 percent counts replies to messages sent after a connection was accepted. Belkins’ 7.2 percent pools campaign types, including connector campaigns that count the request itself as the touch. Worth noting alongside both: Belkins drew the bulk of its touchpoint data from the Expandi platform, so the two are less independent than “two studies agree” would suggest. Treat them as one dataset read two ways.
If you want a single planning number for a message sent to someone who already accepted you, use 10 percent. If you want a number for a cold campaign measured end to end, keep reading, because that number is smaller than either study’s headline.
Is LinkedIn outreach better than cold email?
On reply rate as usually reported, yes, about two to one. On the full funnel, the gap mostly closes. The same Expandi data puts cold email response at 5.1 percent, down from 7 percent the year prior. Against LinkedIn’s 10.4 percent message reply rate, that reads as a clean win for LinkedIn.
The two numbers do not share a denominator.
Work a list of 1,000 people through each channel:
Cold email. All 1,000 get the email. At 5.1 percent, about 51 people respond.
LinkedIn. All 1,000 get a connection request. At 28.5 percent acceptance, 285 accept. Those 285 get a message. At 10.4 percent, about 30 people reply. Add the roughly 3 percent who reply to the connection note itself before accepting, and the total lands near 60.
Sixty versus fifty-one is a real edge, and it is not the two-to-one advantage the headline reply rates imply. It is also slower: the LinkedIn path takes an acceptance step that can sit for days before the message is even eligible to send.
None of that makes LinkedIn a bad channel. It makes the reported reply rate a post-qualification number. LinkedIn’s 10.4 percent is measured on an audience that already raised a hand. Cold email’s 5.1 percent is measured on strangers. Comparing them directly flatters LinkedIn by exactly the width of the acceptance step.
Are LinkedIn reply rates falling?
One part of LinkedIn outreach is falling hard, and the other is not moving at all. Across the twelve months Expandi measured, connection-note reply rates fell from 3.5 percent to 2.2 percent, a 37 percent relative decline. Over the same period, message reply rates held at 10 to 11 percent with no visible erosion. Acceptance held at 28 to 30 percent, apart from a single dip to 25 percent in November 2025 that Expandi flags but cautions against reading too much into from one month alone. Monthly request volume rose from about 1.11 million to 1.25 million, roughly 13 percent more requests landing on the same finite set of people.
Split those two trend lines by what they actually are:
A connection note is unsolicited outreach to a stranger. It is cold email with a character limit. A message sent after acceptance is a conversation with someone who chose to let you in. The first is decaying at 37 percent a year. The second is flat.
Email went through this. Cold email response fell from 7 percent to 5.1 percent in a single year by Expandi’s count, part of a longer slide we tracked in detail here. LinkedIn’s cold surface is now somewhere in the early part of that same slide, with volume growing 13 percent a year to push it along.
The acceptance rate is probably the next number to go. It has held so far because accepting a connection is nearly free and mildly flattering. Once accepted connections reliably turn into pitches, accepting stops being free, and the 28.5 percent starts behaving the way the 3.5 percent did.
What actually moves a LinkedIn reply rate?
Belkins’ breakdown is useful here, because it separates the things teams obsess over from the things that measurably matter.
Personalization still pays, at a known price
Requests carrying a personalized note reply at 8.2 percent versus 5.3 percent without one. That is a meaningful lift and one of the few levers in the data with a clear, sizable effect.
AI-assisted writing barely registers
Campaigns using AI-assisted messaging reply at 7.3 percent against 7.0 percent for those that do not. Three tenths of a point. Whatever advantage AI-written outreach carried when it was rare has been fully priced in by everyone else doing it. This is the clearest signal in the dataset that the differentiator is no longer how the message reads.
Who you are writing to swamps how you wrote it
Reply rates run from 4.2 percent to 10.5 percent by industry, with venture capital, private equity, and staffing at the top. Cut by role instead of industry, HR and talent professionals reply highest at 10.9 percent. The US market comes in at 6.0 percent, below the global average. Choosing a more responsive audience moves the number further than any rewrite.
What does a 2026 outreach sequence look like?
Both channels sit on the same curve, with LinkedIn a few years behind. Moving budget from one to the other buys you time, not a fix.
What the data supports is building the sequence around the acceptance step rather than around the message. Two examples of what that looks like in practice.
A named-account sequence. Fifty target accounts, not five hundred. Connection request with a note referencing something specific and checkable. Once accepted, a message that continues the note rather than restarting the pitch. Then a physical note to the office that references the LinkedIn thread by name, arriving while the conversation is still recent enough to place. The note is doing something the digital touches structurally cannot: it arrives without competing against every other message in the same queue. If you want to see what that note looks like in a live follow-up sequence, we built the sales version here.
A revival sequence. Prospects who accepted the connection months ago and never replied. They are already past the acceptance step, which is the expensive part. A physical note to that list skips the channel where you are one of forty pending messages. For the mechanics of restarting a stalled deal specifically, see when the deal goes dark.
Both sequences assume something the benchmark data makes hard to argue with: the response-rate gap between digital and physical outreach exists largely because physical volume is low, and most teams give up before the touch that would have worked.
The takeaway
LinkedIn is running the same story email ran, a few years behind, and the 13.2 million-request dataset is specific about which part is already unwinding. The unsolicited part lost 37 percent of its reply rate in twelve months. The invited part did not move.
That split is worth carrying past LinkedIn. When reaching a stranger costs the sender nothing, enough people do it that it stops working, and what keeps working tends to be whatever still costs something to send. Worth knowing which of your touches costs you anything.
FAQ
What is a good LinkedIn connection acceptance rate in 2026?
Platform-wide, 28.5 percent, based on Expandi’s analysis of 13,218,869 connection requests sent between May 2025 and April 2026. That figure held stable between 28 and 30 percent for eleven of the twelve months, dipping to 25 percent in November 2025, even as monthly request volume grew about 13 percent. Rates above 30 percent generally indicate a well-targeted list rather than a better-written request.
Is LinkedIn outreach more effective than cold email?
On reply rate as typically reported, LinkedIn leads roughly two to one: 10.4 percent for messages versus 5.1 percent for cold email. The comparison is misleading because LinkedIn’s reply rate is measured only on people who already accepted a connection request. Run 1,000 prospects through both channels end to end and LinkedIn produces about 60 responses to cold email’s 51, a real but much narrower advantage that also takes longer to arrive.
Are LinkedIn reply rates declining?
The cold portion is declining sharply. Reply rates on connection-request notes fell from 3.5 percent in May 2025 to 2.2 percent in April 2026, a 37 percent relative drop in twelve months. Reply rates on messages sent after a connection was accepted stayed flat at 10 to 11 percent over the same period. The decline is concentrated in unsolicited outreach to strangers, which is the part of LinkedIn that most resembles cold email.