LinkedIn closed the door on spray-and-pray. Weekly limits are dynamic, acceptance rate governs your reach, and generic templates barely clear 3%. What still works is narrower targeting and messages that could only have been written to one person.
⏱ 4 min read
LinkedIn outreach is starting conversations with people you don't know, on a platform where they expect to be approached professionally. That expectation is the whole advantage: reply rates on LinkedIn run roughly double cold email. It is also the whole constraint, because the platform now actively polices how you use it.
Three things, all of them limiting volume and rewarding relevance.
| Metric | Typical | Good |
|---|---|---|
| Connection acceptance | 15–30% generic | 40–60% personalised and well targeted |
| Reply rate after acceptance | 10–25% | 25–35% |
| Cold InMail reply | 10–25% | 30–40% |
| Touches before a conversation | ~5 | — |
Benchmarks vary widely by industry and by whoever published them — recruiting sits far above software, where inbox saturation pushes replies down toward 5%. Treat every figure here as directional and measure your own.
At 100 invitations a week, a 20% acceptance rate and a 10% reply rate gives you two conversations. The same 100 at 50% acceptance and 25% reply gives you twelve. Nothing about the volume changed — only the targeting and the writing. That ratio is why every remaining module in this course is about being more specific rather than sending more.
The Only LinkedIn Outreach Video You Will Ever Need — lemlist. A vendor's walkthrough of multichannel sequencing — useful structure, read past the product placement. Open on YouTube ↗
Outreach is one input, not a strategy. It compounds badly on its own and well alongside publishing, because someone who has seen your name in their feed accepts at a far higher rate than a cold stranger. If you are going to do one thing before sending a single request, it is the next module.
It suits B2B, professional services, recruiting, partnerships and anything with a considered buying process. It suits consumer products, low-value transactions and audiences who do not use LinkedIn professionally far less. If your buyers are local homeowners or retail shoppers, the hours spent here are better spent almost anywhere else — and no amount of message craft fixes a channel mismatch.
| Tier | What it adds | When it's worth it |
|---|---|---|
| Free | Basic search, limited noted invites, commercial use limit | Testing whether the channel works at all |
| Premium | InMail credits, who viewed your profile, more search headroom | Light, steady outreach |
| Sales Navigator | Boolean filters, saved lead lists, alerts, monthly InMail credits | Any sustained prospecting motion |
Start free for a fortnight. If you get conversations from fifty well-targeted requests, upgrade. If you get none, the problem is the targeting or the offer, and paying for better filters will not fix either.
⏱ 4 min read
Bad targeting is the root cause of nearly every bad outreach number. If the wrong people receive a perfect message, they still decline.
Your ideal customer profile should be specific enough that you can tell within ten seconds whether a given profile qualifies. Four dimensions usually do it: company type (industry, size, region), role (the specific titles who own the problem), trigger (what makes this relevant now), and disqualifier (who looks right but isn't).
"Marketing managers" is not an ICP. "Marketing managers at 20–200 person B2B firms in Ohio who just posted a job for a web developer" is — and the difference shows up directly in acceptance rate.
| Source | Good for | Cost |
|---|---|---|
| LinkedIn search + filters | Starting out, small volumes | Free, but capped |
| Sales Navigator | Boolean filters, saved lists, lead alerts | Paid, usually necessary at volume |
| People who engaged your posts | The warmest list you own | Free |
| Event and group attendees | A built-in reason to connect | Free |
| Existing clients' networks | Mutual connections lift acceptance | Free |
Free accounts hit the commercial use limit — a monthly cap on searches and profile views that LinkedIn will not disclose, quantify or lift on request, resetting on the first of each month. Prospecting research burns through it quickly, which is the real reason most people doing this seriously end up paying for Sales Navigator.
The highest-responding outreach is timed to something that just happened: a funding round, a leadership change, a job posting that reveals a priority, a post they wrote, a conference they attended, an office opening. A trigger gives you a first line that could not have been written last month, and reported response lifts from trigger-based timing are large — commonly cited around 30% over untimed outreach.
Follow the person. Read their last three posts. Leave one comment that adds something. Then send the request a day or two later. Profile views and prior engagement both raise the odds your name looks familiar rather than unknown — and familiarity is the single cheapest improvement available to acceptance rate.
Spend thirty seconds on each profile before it earns a place on your list: are they still in that role, is the company the size you think, does the problem you solve plausibly sit with them, and is there anything recent worth referencing? People change jobs constantly and stale data is the quiet killer of acceptance rates — a message addressed to someone's job from eighteen months ago reads as automated because it effectively is.
Rank your sources by temperature and work them in order: people who engaged your posts, people in groups you actually participate in, second-degree connections through clients, event attendees, then cold. Most people invert this because the cold list is the easiest to build in bulk — and then wonder why the numbers are bad.
A spreadsheet or CRM with name, company, role, source, trigger, date contacted and outcome. LinkedIn's own interface is a poor system of record: you cannot see your sequence at a glance, you will lose track of who is at which stage, and if the account is ever restricted the history goes with it.
⏱ 4 min read
Every request you send prompts the same action: they click your profile. Whatever is there decides whether they accept. Fixing it takes an hour and improves every message you will ever send.
The Featured section puts three tiles near the top of your profile. Use them for the thing you want a prospect to click: a case study, a service page, a genuinely useful resource. Recommendations from clients do more for credibility than any adjective you can write about yourself — ask three, and give three.
Publishing once or twice a week changes the economics of outreach entirely. Someone who has seen a useful post from you accepts a request from you at a far higher rate, and your profile stops looking like an account that only exists to send messages. It does not need to be clever — what you learned this week, what a client asked, what most people get wrong about your field.
LinkedIn also publishes a Social Selling Index score (find it at linkedin.com/sal/ssi) measuring brand, targeting, engagement and relationships. Do not optimise for it directly; it is a useful rough gauge of account health, and account health is part of what governs your limits.
Open your own profile in a private browsing window, as a stranger would see it, on a phone. Read only what is visible without scrolling or tapping "see more". If that fragment does not tell someone what you do and who you do it for, fix it before you send another request — it is the version most prospects will ever see.
The test: could a prospect read your headline and know whether you solve a problem they have? If not, rewrite it as a plain sentence naming the who and the what.
When someone clicks through from a request, they can see your recent posts and comments. An account with no activity in six months looks dormant; an account whose only activity is promotional posts looks like a broadcast. A handful of genuinely useful posts and thoughtful comments on other people's work does more for acceptance than any About-section rewrite.
Creator-oriented settings change your profile's default button from Connect to Follow, which is right if you want an audience and wrong if you want conversations. If outreach is your priority, make sure connecting is easy and obvious for the people you have just approached.
⏱ 4 min read
You get 300 characters and one shot. Free accounts are limited to a small number of noted invitations per month — widely observed around five — so at any real volume this is a paid-tier activity.
Do not pitch in the connection note. The note's only job is to earn the acceptance. Selling in 300 characters to someone who has never heard of you is the most common mistake in LinkedIn outreach and the fastest route to a declined request and a dented acceptance rate.
Most published tests favour a personalised note, with acceptance around 45% against roughly 15% for a blank request, and a note lifts post-acceptance reply rates as well. A blank request can perform fine when you are clearly relevant — same industry, many mutual connections — and badly when you are not. If you cannot write a note that proves you looked at their profile, that is a signal about your targeting, not your copywriting.
Spread requests across the week — 15 to 20 a day rather than 100 on Monday morning, which is the pattern automation produces. Withdraw stale invitations periodically to keep your pending queue tidy, but know that once withdrawn you cannot re-invite that person for about three weeks, so never withdraw someone you still want.
Weak: "Hi Dana, I'd love to connect and share how our agency helps businesses like yours grow online." Generic, pitches immediately, could have been sent to anyone.
Better: "Hi Dana — saw you're hiring a web developer at Northcoast. I work with agencies on the same stack, always interested to see how teams structure that role. Happy to be connected." Specific, no ask, gives a reason.
Also fine: "Hi Dana — your post on cutting onboarding from six weeks to two was the most useful thing I read this week. Following your posts from now on." Short, honest, and it earns the acceptance on its own merits.
Most people accept and say nothing. That is normal and not a rejection — a connection is a foot in the door, and they will see your posts from now on. Work the sequence in the next module, and if it ends without a reply, leave them connected. A meaningful share of business comes from people who accepted a request eighteen months earlier and got in touch when the need appeared.
⏱ 4 min read
They accepted. This is the moment most outreach dies — either in silence, or in an instant pitch that undoes the goodwill the note earned. Most replies come from follow-ups rather than first messages, and yet roughly half of salespeople never send a second.
| When | Message | Goal |
|---|---|---|
| Day 0–1 | Thanks, plus one line of genuine context. No ask. | Be a person, not a funnel |
| Day 3–5 | Something useful and specific to them — an observation, a resource, a question | Earn the right to a reply |
| Day 8–12 | The soft ask, framed as a question rather than a booking | Open a conversation |
| Day 20–25 | One graceful close-out | Leave the door open |
Two to three follow-ups spread across a week or two is the practical range; spacing messages a few business days apart measurably outperforms one-off sends. Beyond four messages with no response, stop — you are training someone to associate your name with being pestered.
Keep messages under 100 words, three or four sentences, readable in under thirty seconds. LinkedIn's own analysis of tens of millions of InMails found messages under 400 characters outperform average by a wide margin while long ones underperform. Write the way you would message a colleague: no "I hope this message finds you well", no three-paragraph company history, no attachment nobody asked for.
A 60-second voice note inside a warm thread is reported to lift replies substantially over text, because almost nobody sends them and they are obviously not automated. Short video messages work similarly. Both are high-effort and should be reserved for prospects worth the minute.
Treat the goal as a conversation, not a booking. A reply that says "not right now, but ask me in the spring" is a win, and a polite no keeps your acceptance and reply rates healthy in a way that a pushy pitch does not. Outreach where you are visibly willing to take no for an answer converts better than outreach that isn't.
LinkedIn plus email substantially outperforms either alone, because the second channel catches people who never open LinkedIn messages. The sequence that works: connect and message on LinkedIn, then — if you have a legitimate business email — one email referencing the LinkedIn thread rather than pretending it did not happen. Two channels is the ceiling. Adding a phone call to someone who has ignored four messages is not persistence, it is pressure.
Sort responses into three buckets and treat them differently. Interested — move to a conversation quickly, do not make them wait. Not now — note the timing they gave you and actually come back then, which almost nobody does. No — thank them and stop, because a clean exit protects both your reputation and your reply-rate metrics.
The Best LinkedIn Outreach Strategy For 2026 — Sales Feed. A practitioner walkthrough of sequencing and message structure from a B2B sales channel. Open on YouTube ↗
⏱ 4 min read
Two ways to lose an account: send too much, or automate it badly. Both are avoidable.
| Thing | Practical ceiling |
|---|---|
| Connection requests | ~100/week typical; 150–200 for high-trust accounts; far less for new or flagged ones |
| Connection note | 300 characters; free accounts get only a handful of noted invites a month |
| Sales Navigator InMail | A fixed monthly credit allowance; credits are returned when someone replies within 90 days |
| Free account research | Commercial use limit — undisclosed monthly cap on searches and views, resets on the 1st |
| Withdrawn invitation | Cannot re-invite that person for roughly three weeks |
LinkedIn's Prohibited Software and Extensions policy bans unauthorised bots, scrapers and browser extensions that automate activity, and its User Agreement gives it grounds to restrict or remove accounts that use them. Browser-based automation is the most detectable kind. People do use these tools and some get away with it; the risk is your account and your professional presence, on a platform where a restriction is not appealable in any reliable way.
The defensible middle ground: automate the research and the drafting, not the sending. Build lists, gather context, draft variants with AI — then read and send each message yourself. That keeps you inside the rules and, at the volumes LinkedIn now permits, costs less time than the tooling would save.
A spreadsheet is enough: requests sent, accepted, replied, positive replies, meetings. From those you get acceptance rate, reply rate and the only number that matters commercially — conversations per hundred requests. Review weekly, change one variable at a time, and give any change two weeks before judging it.
If acceptance drops below about 30%, stop sending and fix targeting before your limits tighten. If replies dry up while acceptance holds, the problem is your follow-up copy. The two numbers point at different fixes, which is why keeping them separate is worth the spreadsheet.
The response is the same in every case: stop sending entirely for a week or two, keep using the account normally — posting, commenting, reading — and resume at half your previous volume with tighter targeting. Pushing through a warning is how a temporary restriction becomes a permanent one.
Two hours total, split across the week. Monday: build and qualify the list for the week. Tuesday to Friday: 15–20 requests a day and work the follow-up queue, twenty minutes a day. Friday: update the spreadsheet, calculate the two rates, and note one thing to change next week. That cadence is sustainable indefinitely, which matters far more than any single week's volume.
⏱ ~6 min
Answer all ten, then score yourself. Explanations appear after scoring.
All consulted September 2026. Search and AI products change fast — check the primary documentation before acting on anything time-sensitive.