LinkedIn Acceptance Rates Fell From 44% to 32% in Four Months

Every quarter we look at what actually happens to connection requests sent through Banyan. This is the first time we have published it.

The short version: acceptance is falling, steadily, and faster than we expected. Across 6,312 connection requests that have had a full three weeks to be accepted, 2,244 were accepted. That is 35.6% overall. But the overall number hides the trend, and the trend is the story.

Acceptance fell every single month

Month	Requests sent	Accepted	Acceptance rate
April 2026	281	124	44.1%
May 2026	2,191	851	38.8%
June 2026	2,182	746	34.2%
July 2026	1,658	523	31.5%
Total	6,312	2,244	35.6%

Connection requests sent through Banyan, 27 April to 21 July 2026. Cohort-complete: every request counted has had at least 21 days to be accepted.

That is a 12.6 point fall in four months, or a 29% relative decline. It did not bounce in any month.

April is a partial month, covering only the last four days, so treat that 44.1% as the least reliable row in the table. Even dropping April entirely, the trend from May to July is 38.8% to 31.5%, and it is monotonic.

We are not going to pretend we know the single cause. Three explanations are consistent with what we see, and we cannot separate them with this dataset alone:

• The channel is getting more crowded. More tools, more automated outreach, more requests competing for the same inbox.
• Recipients are getting better at ignoring it. A connection request from a stranger reads differently in 2026 than it did in 2023.
• Our own mix changed. More accounts joined over the period, and newer accounts run colder lists than established ones.

The third is the one that should make you most cautious about reading this as a market-wide trend, and we cannot rule it out. Only three accounts sent enough volume in both May and July to support a like-for-like comparison, which is too few to conclude anything from. So the honest version is this: the decline is clear in the aggregate, and we cannot yet prove it is not partly our own mix changing.

Almost every acceptance arrives in the first week

This is the finding we did not expect to be so lopsided. Of every connection request that was eventually accepted:

Accepted within	Share of all acceptances
1 day	71.2%
7 days	95.6%
14 days	98.9%
21 days	99.8%

Time from send to acceptance, measured over the same cohort-complete set as the table above. Requests sent too recently to be cohort-complete are excluded here too, because including them would bias the distribution toward short waits.

Seven in ten of your acceptances land on day one. By the end of week one you have 96% of everything you are ever going to get.

“

If a connection request has not been accepted within seven days, it is not pending. It is declined, quietly.

That has two practical consequences.

Your follow-up timing is probably wrong. If you are waiting two or three weeks before treating a request as dead, you are holding pipeline open on prospects who are gone. Week two contributes 3.3 percentage points of acceptances. Week three contributes 0.9.

Withdrawal windows can be much shorter than most people run them. Pending invitations count against LinkedIn’s limits. If 96% of the value is realised in seven days, holding invitations for 30 days is spending your invitation ceiling on requests that are almost certainly not coming back.

What we would do with this

Three things, in order of how much they matter:

1. Stop measuring acceptance against an old benchmark. If you set your targets when 45% was normal, your campaigns are being judged against a number the channel no longer returns. Around 32% is what a cold-ish B2B list returns right now.
2. Move your follow-up and withdrawal windows in to 7 to 10 days. The tail is not worth waiting for, and pending invitations are a scarce resource.
3. Treat the decline as a reason to warm the list, not to send more. Sending harder into a falling acceptance rate is how accounts get restricted, and every point of decline means each send buys less. We have not published a warm-versus-cold split from this dataset because we cannot segment it cleanly yet, so take this one as advice from running campaigns rather than as a finding from the table above.

Methodology

Anyone should be able to check this, so here is exactly what was counted.

• Source. Connection requests recorded in Banyan’s campaign_actions table with a completed send. A request counts as accepted when an acceptance has been recorded against it.
• Window. Sends between 27 April 2026 and 21 July 2026, analysed on 11 August 2026. The window end moves with the analysis date, because it is always 21 days behind it.
• Cohort completeness. Only requests sent at least 21 days before analysis are included. Recent requests have not had time to be accepted, and including them drags the most recent month artificially low. This is why the report stops at 21 July rather than running to the day of publication, and it is the correction most benchmark posts skip.
• Why 21 days. 99.8% of all acceptances in the dataset arrived within 21 days, so the window captures essentially every acceptance a request will ever get.
• Sample. 6,312 requests, 12 sender accounts, 28 campaigns.
• Suppression. No cell with fewer than 50 requests is reported, and no per-account or per-campaign breakout is published, so no individual customer’s performance can be reconstructed.
• Timing precision. Acceptance times are recorded when Banyan observes the acceptance, not when the recipient tapped the button. Day-level buckets are reliable. We have deliberately not published a median time-to-accept, because at sub-hour resolution it would be measuring our sync interval rather than human behaviour.

Known limits

We would rather state these than have someone find them.

• One platform, and a small number of sender accounts. Twelve accounts is enough to see a trend and not enough to be authoritative.
• The audience skews Indian B2B. Acceptance norms differ by market, and we would expect different absolute numbers elsewhere.
• Account mix changed over the period, which is confounded with the time trend.
• April covers four days, not a month.

Citing this

These figures are free to use with attribution and a link to this page.

• LinkedIn connection request acceptance averaged 35.6% across 6,312 requests sent through Banyan between April and July 2026 (cohort-complete, 21-day window).
• Monthly acceptance fell from 44.1% in April 2026 to 31.5% in July 2026, a 29% relative decline.
• 71.2% of accepted requests were accepted within one day, and 95.6% within seven days.
• Sample: 12 sender accounts, 28 campaigns, India-heavy B2B audience.

If you need a cut we have not published, or want a figure checked before you print it, email hello@getbanyan.co. We will also tell you when we do not know.