In short
- 1"The export is the easy half. The cut from 300 names down to 30 decides everything that follows."
- 2"Write to the idea in the post, never to the act of liking it. \"I saw you liked\" is a receipt, not personalisation."
- 3"Commenters leave text behind, so their list is shorter, richer, and worth messaging before the likers."
- 4"Bursts get accounts restricted more reliably than totals. Thirty qualified names fit inside two ordinary weeks of activity."
- 5"A non-reply is a lead whose next signal has not happened yet. The names that engage twice are the ones to prioritise."
You have the file. Three hundred rows, each one a name, a headline, a profile URL, and the fact that this person reacted to the same post. Getting it took about ninety seconds, and it is where almost every guide on this subject stops.
This one starts there. The distance between an export and a conversation is four decisions: which names to keep, what the first line says when all you know is which post someone reacted to, how fast you are allowed to send, and what happens to the people who read your message and say nothing. Get those wrong and 300 names produce nothing except a slightly worse LinkedIn account.
What to do after you export LinkedIn post likers
Cut the list before you write anything, message the survivors one at a time with a first line about the idea in the post rather than the click, spread the sends across days instead of hours, and give every non-reply exactly one follow-up before moving that person to a watch list. Four steps, in that order.
The order is not decorative. Every step after the first spends part of a limited sending budget, so the cheap step has to carry the weight. Most people invert it, message everyone, then try to fix the results with better copy. Copy cannot fix a list.
Step 1: cut 300 likers to the 30 worth writing to
Run five filters in order, cheapest first, and drop a name the moment one fails. Was the post about the problem you solve, is the role right, is the company right, have you already spoken, and can you write a first line that could not have been sent to anybody else. A tenth of the list surviving is normal, and that is the export doing its job rather than failing at it.
Post relevance runs once, before you look at a single name. A like on a post about hiring tells you nothing about whether someone has the problem you solve. A like on a post about that problem tells you a great deal. This is the only filter that can save you from running the other four, and it is the one people skip because they have already paid for the export emotionally. If the post went wide for reasons unrelated to your offer, delete the file and go search LinkedIn posts by keyword for one that did not.
Role and employer take out most of the rest. Read the headline against the person you actually sell to, not a title you would settle for. Then check the company: right person, wrong company is still a no, and size band, sector and location are all visible on the company's own page.
Prior contact is the one that embarrasses you. A second approach from the same company three weeks after the first reads as a machine, however well it is written.
The first-line test is the last gate, and it is brutal. Try writing the opening sentence. If it would work just as well on any other row in the file, that person has not been qualified, they have been kept. Cut them. The longer version of this cost-ordered method is in intent signals you can read without Sales Navigator.
Likes and comments are not the same list
A like costs one tap and often means "I read the first line". A comment costs visible effort and leaves text behind, so it tells you what the person thinks rather than that they scrolled past. Both come off the same post URL, so pull both. Take the commenters first with the LinkedIn comment scraper, treat them as a separate and much shorter list, and work them before the long tail. Then export the likers of any post and run those names through the same five filters, expecting a lower survival rate.
Ranking the two lists this way is not a preference. It changes what you can write, which is the whole of the next step.
Step 2: the first line when all you know is which post they reacted to
Write to the idea in the post, not to the act of engaging with it. "I saw you liked a post about X" is not personalisation, it is a receipt, and it tells the reader you have a tool. The line that works states the idea, adds one thing you actually think about it, and leaves the person something they could disagree with.
Which version you get depends entirely on what came out of step one.
They liked a post somebody else wrote. The weakest case and the most common. You know one thing: this person was thinking about your topic recently. So write about the topic. One sentence of your own view, one question answerable in a line. Do not mention the post, the like, or how you found them. The relevance works invisibly.
They commented. Their comment is your brief. Quote the claim, not the person. Nadia R., a support lead, writes under a post that her escalation queue doubles at every quarter end. Your opening is already written: refer to the quarter-end spike, say the one thing you believe about where teams try to fix it, and stop.
They engaged with a post you wrote. Now you have standing, and you can say so plainly. This is the only case where naming the post is not creepy, because you were one of the two people in the interaction.
Keep the connection note under LinkedIn's 300-character cap, aiming closer to 200, and the first message after connecting to roughly 75 to 100 words. More shapes, organised by the signal you actually hold, are in the LinkedIn cold message template collection.
Step 3: pace the sends so the account survives the list
Thirty messages is not a morning's work, it is a fortnight's. LinkedIn publishes no official ceiling for connection requests, and the working consensus sits near 100 a week, roughly 15 to 20 on a working day. What gets accounts restricted is rarely the daily total anyway. It is the burst.
Those are budgets, not targets. Thirty qualified people fit comfortably inside two ordinary weeks, which is the real point of step one: pacing stops being a constraint once the list is the right size. Three hundred unqualified names fit inside nothing, which is when people go shopping for a tool that sends faster. That is the wrong end of the problem. How platform limits and tool limits stack is covered in the LinkedIn connection request limits breakdown.
There is a genuine tension between pacing and freshness. Engagement decays, so a message landing two weeks after the post is a cold message wearing a warm coat. Resolve it by ordering, not by speeding up: commenters the day you finish the list, strongest-fit likers the day after, marginal names last or never. If the tail is still unsent when it stops being timely, that is information about the tail.
The discipline is worth it because the trigger is what you spent step one protecting. LeadRiver's April 2026 analysis of more than 50,000 connection requests put trigger-based notes at 50 to 60 percent acceptance against 15 to 25 percent for generic ones.
Step 4: what to do with the ones who do not reply
Send exactly one follow-up, three to five days later, with a different angle rather than a repeat. Then stop messaging and start watching. Someone who ignored one message is not a dead lead, they are a lead whose next signal has not happened yet, and the second signal is worth more than the second message.
"Just floating this back to the top" adds nothing and confirms the first message was a sequence. A second angle on the same problem gives the person a reason to read something they have already skipped once.
Then move the non-repliers to a watch list and re-run the export on the next relevant post a month later. This is where the method compounds, and almost nobody does it: the names that appear on two different posts about the same problem are your real list. One reaction is a scroll. Two reactions to two separate posts on the same subject, weeks apart, is an ongoing interest, and the intersection is short enough to write to properly. That overlap is the most useful thing a liker export produces, and it only exists if you kept the first file.
One thing that looks like failure and is not: an accepted invitation with no reply is still a durable win. That person is now a first-degree connection, and messages to existing connections reply at 12.2 percent against 7.9 percent for cold connect-then-message, in Belkins' 2026 LinkedIn outreach study of 15.1 million touchpoints. Reading their later signals is the subject of signal-based selling on LinkedIn.
Turning LinkedIn comments into calls
A comment becomes a call in three messages, not one. The first refers to the idea in their comment and asks a question they can answer in a sentence, with no ask attached. The second answers back with something useful whether or not they ever buy. Only the third asks for time, and it names what the fifteen minutes would cover.
The mistake is asking in message one. A comment proves interest in a topic and nothing about interest in you, so a calendar link in the opener asks the person to cover that whole distance in one step.
- Message one: earn a reply, not a meeting. One position on the thing they said, one question that costs them ten seconds.
- Message two: be useful with no strings. Tell them the thing you know that they do not. This decides whether the third message gets read.
- Message three: ask for a specific short slot. "Worth fifteen minutes on the quarter-end spike" beats "open to a quick chat", because it says what they are agreeing to.
When they reply but decline the call, take the reply as the result. Keep answering in the thread and let the call happen when it is obviously the faster medium. A no to a meeting is not a no to a conversation.
The same four steps, without the spreadsheet
All of this works by hand, and for one post a month it should be done by hand. It stops scaling when the filtering, the writing and the pacing have to happen every week, on lists nobody had time to read.
That is the loop BeReach runs. It finds people who just showed interest on LinkedIn, checks them against who you actually sell to, writes one message per person rather than one template with a name slot, and sends from your own account at a paced rate, with every send approved by you first. It is not a scraper and it is not unattended automation, and the difference matters most on the step this article is about: the cut is a judgement call, and the agent makes it explicit so you can overrule it.
On the free tools, the privacy position is worth stating plainly, because it comes up: we do not store your credentials or your data. Your session is only passed through to run your own search, and the leads go straight to you.
Every viral post is 100+ warm conversations waiting.
Tell your agent who you want to reach. It finds them, says which ones are worth your time, writes the first line, and follows up.
Frequently asked questions
How do you scrape LinkedIn likes for outreach?
Export the reaction list from one relevant post, then filter it before writing anything. Cut by post relevance, role, employer and prior contact, in that order, cheapest checks first, since a skipped filter shows up later as a message that should never have gone out. Message the survivors individually with a first line about the idea in the post rather than the fact that they liked it, spread across days rather than hours.
How many likers from one post are actually worth messaging?
Roughly a tenth on a well-chosen post, and far fewer on a post that went wide for reasons unrelated to your offer. Three hundred reactions filtered down to thirty people who match the role, the company profile and the topic is a normal result. Optimising for a longer surviving list is how engagement exports become spreadsheets nobody opens.
Should I message post likers or commenters first?
Commenters, in almost every case. A comment costs visible effort and leaves text behind, so you get a stronger intent signal and a ready-made opening line. Liker lists are longer and weaker. Pull both from the same post URL, work the commenters the day you export them, and treat the likers as the long tail behind them.
How fast can I message people from a liker export?
Slower than the export arrives. LinkedIn publishes no official ceiling for connection requests and the working consensus is near 100 a week, roughly 15 to 20 a working day. Bursts trip restrictions more reliably than totals, so space the sends by minutes and the list by days. Thirty qualified people fit inside two ordinary weeks of activity.
How do I turn LinkedIn comments into sales calls?
In three messages. The first takes a position on the idea in their comment and asks one question that costs them ten seconds. The second answers with something genuinely useful and asks for nothing. The third asks for a specific short slot and names the topic. Asking for the call in the opening message covers the whole distance at once, which is why it rarely works.
Run the cut, not the blast. BeReach finds the people who just engaged with a post in your space, grades them against who you actually sell to, drafts one message per person, and paces the sends from your own account with every message approved by you first. Pro is EUR 89 per month billed quarterly, EUR 99 at the regular rate, with a 3-day free trial. See how BeReach runs your outreach.
Reading this in an AI assistant? Hand it the page and let it summarize, so you can ask follow-up questions against the whole argument rather than the part you have read so far.
