What are ethical LinkedIn outreach guidelines?

They are the rules that keep outreach honest toward the person receiving it: contact people you have a real reason to contact, say who you are and why, personalise only from what they published, respect silence, and keep volume low enough that the messages stay individual. This page states 9 of them, and 6 carry a measured figure from Reachium's own LinkedIn outreach data rather than an opinion.

How many LinkedIn connection requests per day is reasonable?

25 is the practical ceiling: across 10,023 active sending days the median day and the 90th-percentile day are both exactly 25 invites, and fewer than 1% of days exceeded it. Acceptance peaks lower still. On a separate frozen cohort of 180,155 matured requests, accounts averaging 10 to 19 invites a day were accepted 32.0% of the time against 26.7% for accounts averaging 20 to 29. Guidance recommending 50 to 70 requests a day does not match what sending accounts actually do.

Is LinkedIn automation against LinkedIn terms of service?

LinkedIn's User Agreement prohibits using software to scrape or to automate activity on the platform without permission, and it applies to every tool in this category, including Reachium. Vendors who tell you their product sits outside that are describing marketing, not the agreement. What differs between tools is how much risk a given sending pattern carries and how much control you have over volume and behaviour, not whether the terms apply.

Can LinkedIn outreach get my account restricted?

Yes. Any automated activity carries a risk of restriction, and no tool or sending pattern removes it. The risk is managed rather than eliminated: lower daily volume, human-paced sending, relevant targeting and clean stopping rules all reduce how much unusual activity an account shows. Anyone promising your account cannot be restricted is promising something they do not control.

Does personalising a connection request actually work?

It changes the conversation more than the connection. Among people who accepted a request, 47.8% replied when it carried a note against 25.5% when it did not, measured across 19,516 noted and 160,639 blank requests. Whether a note changes acceptance was tested in the same study and came out inconclusive, so the defensible claim is that a note earns you a better conversation, not an easier connection.

Is it ethical to send automated LinkedIn messages at all?

Automation is a scheduling question, not an ethical one. What makes outreach defensible is whether the recipient was chosen for a real reason, told who is writing and why, personalised only from what they published, and left alone when they do not answer. A hand-typed message that fails those tests is worse than a scheduled one that passes them.

Guidelines

Ethical LinkedIn Outreach Guidelines

Most guidance on ethical outreach is a list of opinions with a number attached to none of them. 9 rules follow. 6 of them state a figure measured from Reachium’s own LinkedIn sending data, and each links to the study that measured it, with its sample and its limitations. The remaining 3 are principles, and are marked as principles.

The short version

01Treat 25 invites a day as the ceiling, because it already is one
02Sending less is not only safer, it performs better
03Personalise for the conversation, not for the acceptance rate
04Design for the people who will not reply
05Do not react to a bad week
06Assume the inbox is more crowded than it was
07Personalise only from what the person chose to publish
08Say who you are and why you are writing, in the first message
09Treat silence as a no, and a no as permanent

The rules

Treat 25 invites a day as the ceiling, because it already is one

Plan campaigns around a maximum of 25 connection invites per account per day. Advice to send 50, 70 or 100 a day describes something that does not happen in practice.

median and p90 invites on an active sending day

Across 10,023 active sending days, the median day and the 90th-percentile day are the same number: 25 invites. 75.8% of active days ran at 25 or more. Fewer than 1% of days recorded more than 25, and those rows are send-attribution artifacts rather than a way past the platform ceiling.

Method and sample: Everyone Runs at the Cap

Sending less is not only safer, it performs better

Target 10 to 19 invites per account per day. The restraint is not a tax you pay for being careful, it is where acceptance peaked.

32.0% vs 26.7%

acceptance at 10-19/day vs 20-29/day

On a frozen cohort of 180,155 matured requests, acceptance peaked at 32.0% for accounts averaging 10 to 19 invites a day and fell to 26.7% for accounts averaging 20 to 29. The 20 to 29 band carried 89.6% of the requests measured, so most senders sit in the worse-performing band by default.

Method and sample: The Volume Tax

Personalise for the conversation, not for the acceptance rate

Write a note that refers to something the person actually published, and judge it on replies rather than on acceptance rate.

47.8% vs 25.5%

reply rate of accepted, noted vs blank

Among people who accepted a request, 47.8% replied when the request carried a note against 25.5% when it did not, measured across 19,516 noted and 160,639 blank requests. The same study tested whether a note changes acceptance and found the result inconclusive, so the honest claim is narrower than the usual one: a note is not reliably shown to get you connected, it is shown to change what happens once you are.

Method and sample: Do Connection Notes Work

Design for the people who will not reply

Build the sequence so a non-response costs the recipient nothing and ends cleanly. Non-response is the normal outcome, not a failure state to engineer around.

27.1%

connection acceptance rate

Across 180,155 matured connection requests, 27.1% were accepted, 27.6% of accepted connections produced a reply, and 2.0% of accepted connections produced a booked meeting. A well-run campaign is one where the large majority of recipients quietly ignore it.

Method and sample: LinkedIn Outreach Benchmarks 2026

Do not react to a bad week

Change one variable at a time and give it a full month. Reading short-run movement as signal is what drives the escalation that puts accounts at risk.

5.69 pts

median month-over-month acceptance swing

An analysis of 82 same-workspace consecutive-month pairs found the median month-over-month acceptance swing was 5.69 points, with 54.9% of pairs swinging 5 points or more and 31.7% swinging 10 or more. That is the amount a rate moves without anyone changing anything.

Method and sample: The Noise Floor

Assume the inbox is more crowded than it was

Raise the bar for who is worth contacting every quarter. The same message to the same list earns fewer replies than it did a year ago.

32.2% to 22.0%

reply rate of accepted, 2025 to 2026

Holding maturity constant, acceptance rose slightly from 26.3% in 2025 to 28.1% in 2026 while the share of accepted connections producing a reply fell from 32.2% to 22.0%. Getting connected became no harder. Getting a conversation became considerably harder.

Method and sample: The Reply Rate Collapse

Personalise only from what the person chose to publish

Use their posts, their headline, their company page, their public work. Do not use data they did not put on the profile, and do not imply you know more about them than a reader of their profile would.

No measurement

This one is a principle, not a finding. Reachium has no data that tests it, and it is listed separately for that reason.

Say who you are and why you are writing, in the first message

Name the company and the reason for contact before the first ask. A message that hides its purpose until message three is not a nurture sequence, it is a pretext.

No measurement

This one is a principle, not a finding. Reachium has no data that tests it, and it is listed separately for that reason.

Treat silence as a no, and a no as permanent

End the sequence after the planned follow-ups whether or not the person replied, and remove anyone who declines from every future campaign, not just the current one.

No measurement

This one is a principle, not a finding. Reachium has no data that tests it, and it is listed separately for that reason.

What this data cannot tell you

Every figure above comes from campaigns run through one platform, by customers who chose to use it. That is a specific population, not LinkedIn as a whole. Outreach sent by hand, from other tools, or in other markets may behave differently.

The dataset is also concentrated: a small number of workspaces account for most of the requests measured, so a single large sender can move a rate. Each linked study states its own concentration and its own limitations, and several of them report a result that did not reach a conclusion rather than rounding it into one.

None of this measures whether an account was restricted, because Reachium does not observe LinkedIn’s enforcement decisions. Lower volume and human-paced sending reduce how much unusual activity an account shows. They do not put an account outside the reach of enforcement, and no tool or sending pattern does.

Questions

What are ethical LinkedIn outreach guidelines?: They are the rules that keep outreach honest toward the person receiving it: contact people you have a real reason to contact, say who you are and why, personalise only from what they published, respect silence, and keep volume low enough that the messages stay individual. This page states 9 of them, and 6 carry a measured figure from Reachium's own LinkedIn outreach data rather than an opinion.
How many LinkedIn connection requests per day is reasonable?: 25 is the practical ceiling: across 10,023 active sending days the median day and the 90th-percentile day are both exactly 25 invites, and fewer than 1% of days exceeded it. Acceptance peaks lower still. On a separate frozen cohort of 180,155 matured requests, accounts averaging 10 to 19 invites a day were accepted 32.0% of the time against 26.7% for accounts averaging 20 to 29. Guidance recommending 50 to 70 requests a day does not match what sending accounts actually do.
Is LinkedIn automation against LinkedIn terms of service?: LinkedIn's User Agreement prohibits using software to scrape or to automate activity on the platform without permission, and it applies to every tool in this category, including Reachium. Vendors who tell you their product sits outside that are describing marketing, not the agreement. What differs between tools is how much risk a given sending pattern carries and how much control you have over volume and behaviour, not whether the terms apply.
Can LinkedIn outreach get my account restricted?: Yes. Any automated activity carries a risk of restriction, and no tool or sending pattern removes it. The risk is managed rather than eliminated: lower daily volume, human-paced sending, relevant targeting and clean stopping rules all reduce how much unusual activity an account shows. Anyone promising your account cannot be restricted is promising something they do not control.
Does personalising a connection request actually work?: It changes the conversation more than the connection. Among people who accepted a request, 47.8% replied when it carried a note against 25.5% when it did not, measured across 19,516 noted and 160,639 blank requests. Whether a note changes acceptance was tested in the same study and came out inconclusive, so the defensible claim is that a note earns you a better conversation, not an easier connection.
Is it ethical to send automated LinkedIn messages at all?: Automation is a scheduling question, not an ethical one. What makes outreach defensible is whether the recipient was chosen for a real reason, told who is writing and why, personalised only from what they published, and left alone when they do not answer. A hand-typed message that fails those tests is worse than a scheduled one that passes them.
What acceptance rate should I expect from ethical outreach?: Across 180,155 LinkedIn connection requests given at least 30 days to mature, 27.1% were accepted, 27.6% of accepted connections produced a reply, and 2.0% of accepted connections produced a booked meeting. Expect wide movement month to month: inside the same workspace, the median swing between consecutive months was 5.69 points with no change in approach.

Where the numbers come from

Figures on this page are drawn from Reachium’s first-party outreach data: 10,023 active account-days of sending, and matured connection-request cohorts of 196,696 and 180,155 requests frozen at different dates. A request counts as matured once it has had at least 30 days to be accepted. The two cohorts are reported separately and never combined inside one calculation. Each rule links to the study carrying that rule’s sample, window and denominators.

All first-party research