How I do cold outreach on LinkedIn

Hi, I've just tryed the free version.
From my point of view, I've spent only few minutes on the software, but I could say to you to ask a feature in the initial research.
In my case I want leads from Italy (because my Saas now it is only in italian language), but I get lead from every where.

Maybe you could add this select box in the initial page.

thank you

This is really great insight!

This is gold. Point #1 and #4 hit hard.

Quality > Volume and Profile actually matters more than the message.

I'm building AI Agents for BD businesses and struggling with LinkedIn outreach.

Question: When you research those 20 prospects, what's the 1 thing you look for to personalize the first message?

The finding that requests without a note performed about 25% better is especially interesting. I’ve been assuming that a short, personalized note would build more trust, but it may also make the request feel like an immediate sales pitch.

I’d be curious about the metric behind the comparison: did no-note requests only get a higher acceptance rate, or did they also produce more replies and qualified conversations afterward? Acceptance can become another vanity metric if the connection never turns into a real exchange.

LinkedIn outreach takes a lot of time. Have you tried combining your outreach data with automated multi-channel follow-ups? I recently hooked up an automated notification system via Pabbly that syncs e-com actions directly to custom WhatsApp webhooks. Streamlining the backend makes outreach so much easier.

The hardest part of LinkedIn outreach for me is finding the balance between personalization and actually being able to do it at scale. Too generic and nobody replies, too personalized and it takes forever. Interested to see how you structure your first message

Cold outreach is one of those things that looks simple until you actually try to get consistent replies without sounding automated. The biggest difference for me has always been whether the first message feels like a real conversation or a sales pitch. Curious what part of your process had the biggest impact on reply rates

The data point on connection requests without a note is something I had to learn the hard way. Spent about 3 months adding a note to every single request because it felt more professional, before I actually ran the test. Same pattern as what you found.

The reason makes sense once you see it: the note implies you're pitching before you've earned the right to. People check the profile first anyway, so let the profile do the selling.

The ICP specificity point is where most people go wrong before they ever get to message copy. Watched teams spend weeks on a perfect message when the list was the actual problem. Right people at the right time, an average message still converts.

What's your experience been with engaging existing connections vs new outreach? In my experience the second bucket converts at 3-4x but almost nobody works it deliberately.

The data point on connection requests without a note is something I had to learn the hard way. Spent about 3 months adding a note to every single request because it felt more professional, before I actually ran the test. Same pattern as what you found — and the reason makes sense once you see it. The note implies you're pitching before you've earned the right to. People check the profile first anyway, so let the profile do the selling.

The ICP specificity point is where most people go wrong before they ever get to message copy. I've watched teams spend weeks on a "perfect message" when the list was the actual problem. If you're reaching out to the right people at the right time, an average message still converts.

One thing I'd add: what's your experience been with engaging existing connections versus new outreach? In my experience the second bucket converts at 3-4x but almost nobody works it deliberately.

The data point on connection requests without a note is something I had to learn the hard way. Spent about 3 months adding a note to every single request because it felt more professional, before I actually ran the test. Same pattern as what you found — and the reason makes sense once you see it. The note implies you're pitching before you've earned the right to. People check the profile first anyway, so let the profile do the selling.

The ICP specificity point is where most people go wrong before they ever get to message copy. I've watched teams spend weeks on a "perfect message" when the list was the actual problem. If you're reaching out to the right people at the right time, an average message still converts.

One thing I'd add: what's your experience been with engaging existing connections versus new outreach? In my experience the second bucket converts at 3-4x but almost nobody works it deliberately.

One thing that's worked well for me is creating genuinely useful content before doing outreach. When the landing page solves a very specific problem, reply rates improve because you're sharing something valuable instead of just asking for attention. Curious—what kind of reply rate are you seeing from this approach?

The missing layer is measurement at each handoff. Track connection acceptance, reply rate, qualified conversations, and booked calls separately. If acceptance is low, fix targeting or the profile. If replies are low, fix the opener. If calls are low, fix qualification and the offer. That makes outreach a controlled learning loop instead of a volume contest.

Cold outreach is such a weird balance. Personalize too much and it takes forever, automate too much and everyone can tell 😅 What part of your process made the biggest difference in reply rates?

My problem is that I have an old LinkedIn account with approximately 2000 followers.
But none of my connections are my ICP, they're just local random people, connections of connections.

I am struggling to get out of that hole even if im trying to connect with my ICP, LinkedIn is still hardly suggesting not relevant posts.

The "no universal script" one is the myth people resist hardest, and I think that's because a script is a way of avoiding the actual work.

The only personalisation that survives contact is the kind that can't be faked - something you could only know by having actually used the thing. "Congrats on the launch" and "saw you're in the X space" are mail merges with extra steps, and everyone can smell it. Compare that with: I opened your signup form on a phone with a Lithuanian keyboard and the price field rejects a comma decimal. You can't template that, and the recipient knows it in one second, which is the entire point - it proves you spent your time before asking for theirs.

The cost is exactly the 20-vs-200 trade you described. Done properly that's maybe fifteen minutes a prospect, so twenty of them is most of a working day, which sounds terrible right up until you compare it with the reply rate on 200 messages people have already seen three times this week.

A question, since you ran the agency: where does specificity start working against you? I'd assume there's a line past which "here's a concrete flaw in your product" reads as criticism rather than homework, and my guess is it depends more on how the ask is framed than on the finding itself - but you'd have seen the pattern across far more sends than I have. Did the sharp ones convert better, or just get remembered?

The 'no note' stat is the one that surprises people the most, but it tracks — a blank request feels lower-pressure, and the curious prospect checks your profile instead. Also completely agree on pending invites killing your reach. I clean mine weekly and the algorithm rewards it. Solid breakdown.

Effective LinkedIn cold outreach is less about sending a large number of messages and more about reaching the right people with relevant communication. Start by clearly defining your ideal customer profile (ICP), including their industry, job role, company size, location, and specific needs.

Personalization is also important. Instead of relying on a single message template, research each prospect and explain why your offer may be relevant to them. A professional and complete LinkedIn profile can also improve credibility because prospects often check your profile before responding.

Automation can save time, but excessive automation and generic messages can reduce engagement. It is better to focus on a smaller number of well-targeted prospects and build genuine professional relationships.

Finally, maintain a clean and relevant network by reviewing pending invitations and connections regularly. LinkedIn outreach works best when it combines clear targeting, useful communication, personalization, and consistent relationship building rather than simply focusing on connection numbers.

Completely agree with #1 and #2. In the age of AI-generated spam, blast tools have made inboxes almost unusable. Doubling down on a razor-sharp ICP and reaching out with actual context beats 1,000 generic touches every single time. Quality > quantity is the only sustainable way forward.

thanks for sharing this I never realized why pending invites should be withdrawn! I am struggling to draw engagements on my posts despite so many connections.

This is useful — I've been thinking about how to get in front of freelance clients without cold DMs feeling spammy. Do you find LinkedIn outreach works differently for service-based offers versus SaaS, or is the approach basically the same?

The biggest takeaway for me is quality > volume.

I’d add one more: even a highly personalized first message can go nowhere if there’s no thoughtful follow-up behind it.

Getting the right person to notice you is step one. Managing the conversation after that is where most systems break.

The no-note finding matches my experience: a note turns a connection request into a pitch, and a pitch invites a no. What I'd add after years of posting daily on LinkedIn is that your last five posts are the real cover letter, because people check your activity before they accept. Targeting plus a credible profile beats any message template.

The point about quality over quantity is probably the most important takeaway here.

Too many people focus on sending more messages instead of understanding the people they're trying to reach. A well-researched list of prospects will almost always outperform a large list built without clear targeting.

Great insights.

The best outreach usually feels relevant because it’s based on who you’re talking to, not because you found the perfect script.

Great breakdown of what actually matters. 👏

I would love to work with you

I think I may have found the kind of founder I’d actually enjoy working with.

I’ve been reading through your IbexAI journey, and one thing kept standing out to me.

You seem to be doing a little bit of everything.

From figuring out how to get the first paying users, to getting 42 beta testers in two weeks, to writing about finding high-intent prospects, testing distribution channels, and even putting together that list of 60+ subreddits for founders.

And honestly, I thought: someone has to be making sure all the things around the founder don't start becoming another full-time job.

That’s where I think I could fit.

I’m an Executive Assistant, but not the “calendar and reminders” kind of EA.

I’m the person who likes taking the loose ends off a founder’s plate keeping things organized, following up on people and opportunities, handling research, staying on top of moving pieces, supporting customers when needed, and making sure things actually move from “we should do this” to “this is done.”

I know you’re not currently advertising for an EA, and that’s actually why I’m writing to you.

I’d rather start the conversation before there’s a job description telling you what you need.

So, Matt…

When do I start?