TL;DR: Most LinkedIn acceptances come fast: 46.33% within a day and 91.03% within 30 days. An invite still pending at day 30 had a 3.24% chance of ever being accepted, and at day 60 just 0.99%.
Across 189,692 LinkedIn connection requests given at least 120 days to resolve, 91.03% of all eventual acceptances arrived within 30 days of the invite, and of invites still pending at day 30 only 3.24% were ever accepted later. Platform data
Key Takeaways
- The median accepted invite was accepted about one day after it was sent.
- 46.33% of eventual acceptances arrived within one day, 75.28% within a week and 91.03% within 30 days.
- Of invites still pending at day 30, 3.24% were accepted later; at day 60 the figure was 0.99%.
- Acceptances after day 30 added up to 2.44% of all requests sent.
- 72.82% of requests were never accepted, a group that includes invites the sender withdrew.
What did we measure?
Every LinkedIn prospector ends up with a pile of pending invitations, and two questions follow: how long is it worth waiting, and when does an old invite stop being worth anything? Advice on this usually arrives as a rule of thumb with no data behind it.
Reachium's data records when each connection request was sent and, if it was accepted, when the acceptance was synced. This study measures how acceptances spread out over time, and, for invites still unanswered at a given age, how many were ever accepted afterwards. It also reports what the platform's newer withdrawal logs show, within what they can support.
This study makes no claim about any LinkedIn limit on pending invitations. It measures timing, not platform rules.
How was it measured?
Dataset window. Connection requests sent between 8 January 2025 and 16 October 2026, pulled 16 November 2026, matured at least 30 days. For this study the cohort is narrowed to requests sent between 8 January 2025 and 18 July 2026, so every request had at least 120 days before the pull for a late acceptance to arrive. That cohort is 189,692 requests. InMail sends are excluded. Platform data
Timing. Days from send to acceptance. Acceptance timestamps come from a periodic sync, not a live event, so sub-day timing is approximate; day-level figures are reliable. 810 acceptances carry a timestamp at or before the send timestamp, a sync artifact. They are counted in the within-one-day bucket and make up 1.57% of acceptances (derived), too few to move any figure below. Platform data
Two denominators, kept separate. "Share of acceptances within N days" divides by all eventual acceptances (51,558). "Share later accepted" divides by the invites still unaccepted at day N. They answer different questions: how fast yes arrives, and how likely a yes still is.
"Never accepted" means still pending, ignored, or withdrawn by the sender. The data does not separate those three.
What did we find?
Of the 189,692 requests, 51,558 (27.18%) were accepted at some point, and the other 72.82% never were. The acceptances that did come arrived quickly.
| Measure | Value |
|---|---|
| Accepted, ever | 51,558 (27.18%) |
| Never accepted (pending, withdrawn or ignored) | 72.82% |
| Median days from send to acceptance | 1.07 |
| 75th percentile, days | 6.89 |
| 90th percentile, days | 26.91 |
| 95th percentile, days | 44.34 |
How fast acceptances arrive. Nearly half land within a day. The curve flattens sharply after the first week.
| Within | Share of all eventual acceptances |
|---|---|
| 1 day | 46.33% |
| 3 days | 64.87% |
| 7 days | 75.28% |
| 14 days | 82.70% |
| 21 days | 86.91% |
| 30 days | 91.03% |
| 60 days | 97.31% |
| 90 days | 98.69% |
How much an old invite is still worth. This is the question behind "should I withdraw pending connection requests". For invites still unaccepted at each age, here is how many were accepted later.
| Still pending at day | Invites still pending | Later accepted | Share later accepted |
|---|---|---|---|
| 7 | 150,880 | 12,746 | 8.45% |
| 14 | 147,056 | 8,922 | 6.07% |
| 21 | 144,884 | 6,750 | 4.66% |
| 30 | 142,760 | 4,626 | 3.24% |
| 60 | 139,522 | 1,388 | 0.99% |
At day 30, 142,760 invites were still pending, 75.26% of everything sent (derived). Of those, 4,626 were eventually accepted. Those late acceptances are 8.97% of all acceptances and 2.44% of all requests sent. By day 60 the chance of a pending invite ever being accepted is under one in a hundred. Platform data
What withdrawal logs show. Withdrawal logging on the platform is recent: a daily revoke counter began on 30 September 2026 and a withdraw-attempt log on 19 October 2026. In the revoke-counter window, accounts logged 81.76 withdrawals for every 100 invites sent (derived). In the attempt log, 99.90% of withdraw attempts completed (derived) and the rest returned an uncertain outcome. In other words, withdrawing invites is routine practice among these senders, not an edge case. Platform data
The withdrawal log does not link back to the original request, so it cannot show how old an invite was when it was withdrawn. This study publishes no figure for that.
What does this mean for anyone managing a pending pile?
Pending invitations decay fast. A request sent today is most likely to be accepted in the first day, very likely within the first week if at all, and after a month the odds are small. The 3.24% figure is not zero, so an old invite is not worthless, but nine in ten of the acceptances a campaign will ever get have already arrived by then.
That gives the pile a shape. Almost everything sitting in it after 30 days is a request that will never be accepted: 96.76% of day-30 pending invites stayed unaccepted (derived). Whether to withdraw them depends on what a sender values in the slot, and this data does not settle that. It does say the expected cost of withdrawing at day 30 is small: at least about 3 acceptances given up per 100 withdrawn invites (derived from 3.24%, a floor, because some pending invites had already been withdrawn).
The timing also fits the rest of the funnel. Replies to accepted connections cluster fast, and follow-ups convert a modest share of the silent. A companion study of the weekly sending pattern shows how quickly a pile accumulates when accounts send close to every day, and most accounts run at the daily cap. Which plan the sender pays for is linked to the acceptance level; this study does not split timing by plan. Teams planning campaign volume around these curves can compare setups on Reachium's pricing page.
How do you use this in practice?
An SDR looking at an invite sent three weeks ago is holding one from the 144,884-invite group still pending at day 21, of which 4.66% were accepted later. A week later, at day 30, the same invite sits in the group where that figure is 3.24%. Platform data
A team lead deciding when a campaign's results are final can see that 75.28% of acceptances arrive within 7 days and 91.03% within 30. A campaign read at day 7 is missing about a quarter of its eventual acceptances (derived); one read at day 30 is missing under a tenth. Platform data
A RevOps analyst comparing a campaign's acceptance against benchmarks is looking at the same effect from the other side. A rate measured on requests younger than 30 days will understate the final figure. Rates on requests matured 30 days, the standard used across this series, capture about 91% of eventual acceptances (derived from 91.03%). Our analysis
The data cannot tell any of them what an open invite slot is worth to them. It can only price the acceptances they would forgo. The Aera Capital case study shows one team's outreach numbers on the same platform.
What are the limits of this study?
No LinkedIn limit is measured or claimed. This study is about acceptance timing. It says nothing about how many pending invitations LinkedIn allows.
Sync timing. Acceptance timestamps come from a periodic sync. Day-level shares are reliable; anything finer is not. The 810 pre-send timestamps (1.57% of acceptances, derived) are counted in the first day.
Withdrawals hide in "never accepted". Some invites in the pending tables were withdrawn by the sender before they could be accepted. Their true late-acceptance chance is unknown, so the "later accepted" shares are a floor for invites left alone, not a precise estimate. Platform data
Withdrawal logs are short and unlinked. The counter covers autumn 2026 only, and age at withdrawal cannot be measured.
120-day horizon. An acceptance arriving after 120 days is outside the window. The 90-day share (98.69%) suggests very few are missed.
Observational. Targeting, copy and sender behaviour vary across campaigns and are not held constant. Teams that want help running a campaign end to end can talk to the agency team.
FAQ
How long does it take for a LinkedIn connection request to be accepted?
The median was about one day across 51,558 acceptances. 46.33% arrived within a day and 75.28% within a week.
When should you withdraw pending LinkedIn invitations?
The data does not set a rule, but it prices the choice: of invites still pending at day 30, only 3.24% were ever accepted, and at day 60 just 0.99%.
Should I withdraw pending connection requests on LinkedIn?
Withdrawing a 30-day-old invite gives up at least about 3 acceptances per 100 withdrawn, on this data. Accounts here logged 81.76 withdrawals for every 100 invites sent during the logging window, so it is common practice.
Is there a LinkedIn pending invitations limit?
This study does not measure or state one. It measures how quickly acceptances arrive and how rarely old invites are accepted.