Hi friends - This is a revamped piece from my 2024 vintage. I wrote the original in January 2024, after Google and Yahoo announced their new bulk sender rules, and I went back to it this week expecting to make small edits.
The core argument held up completely. The numbers and the tools have shifted.
So here’s the updated version — where outbound actually sits, what makes it perform better, and how to run it on LinkedIn. Enjoy!
The LinkedIn outbound advantage
Before I make the case for LinkedIn, I want to talk about the category it sits in.
At a 50k foot level, there are four ways to build pipeline: inbound, outbound, nearbound and channel, and organic and existing customers. Cold outbound is the lowest-quality of the four, converting at roughly 1%. Inbound produces the best quality leads but takes real time to build. Partnership and nearbound leads come second. Outbound’s genuine advantage isn’t lead quality — it’s speed to activation.
That’s why early-stage teams reach for it first, and it’s a legitimate reason. You don’t need an audience, a partner ecosystem, or a customer base to start. You need a list and a message. When someone says “we need pipeline,” outbound is the one you launch this afternoon.
So run it. Just don’t mistake it for a full GTM strategy, and invest in the higher-quality categories in parallel.
Here’s the part that matters for the rest of this piece: treat the 1% as a ceiling you raise by getting more precise, not by sending more. There are only two levers that reliably move it.
Lever one: build campaigns from signals, not filters. Warm outbound built on a named trigger meaningfully outperforms generic cold.
Lever two: use LinkedIn, not just email. The channel you choose changes your response rate before you write a word.
This article is about lever two, with lever one woven through — because the two compound, and neither works as well alone.
📘 If you want my complete outbound playbook — email deliverability, domain warm-up, subject line research, cadence by segment, and the full tool comparison — it’s all in the cold outbound best practices guide. This article is the LinkedIn deep-dive that guide points to.
LinkedIn vs. email: why LinkedIn wins message for message
Let’s start with the durable part, because this hasn’t changed at all.
When you send a cold message on LinkedIn, the prospect can safely and easily vet you. They see your mutual connections. They see where you’ve worked. They can click into your company. They see your face — and knowing you’re a real human with real experience does a lot of quiet work.
With email, none of that is instantly available to them. You could put a photo or a link in your signature, but both can hurt deliverability. They could look you up but that takes effort most people don’t give a stranger.
And here’s the piece that’s gotten more true since 2024, not less: nobody clicks links in cold emails anymore. Phishing and business email compromise have trained every buyer in your ICP to treat an unexpected link from an unknown sender as a threat. Plenty of them work somewhere with a security team actively coaching them not to click, or with IT tooling that strips and sandboxes the link before it ever reaches them.
So in email, the one action you want — click this, look at us — is the exact action your prospect has been trained to refuse.
On LinkedIn, clicking is safe. Tapping someone’s profile inside the platform carries no risk, and buyers know it. That’s the whole advantage in one sentence: on LinkedIn, a stranger can verify you in about eight seconds, safely. In an inbox, they can’t and they won’t.
The data backs it up. Post-connection LinkedIn messages average a 10.4% reply rate across industries, per Expandi’s analysis of 13.2 million connection requests. Cold email averages 3.4–5.1%. Roughly double, per message.
To be clear, this isn’t LinkedIn instead of email — it’s LinkedIn first, then email. The combination outperforms either channel alone, and it solves a sequencing problem: LinkedIn works from day one while your sending domain is still warming up, so you’re not sitting idle for two months waiting on infrastructure.
What changed since 2024
Three things, and they pull in opposite directions.
Cold email got harder to do badly
The Google and Yahoo rules I wrote about in 2024 stuck. Then Microsoft joined them. Since May 5, 2025, anyone sending 5,000+ messages a day to Outlook, Hotmail, or Live addresses has to pass SPF, DKIM, and DMARC — and non-compliant mail gets rejected outright, not just filtered to junk.
I said at the time these changes were mostly a good thing, and I still think so. They didn’t kill cold email. They killed lazy cold email, and they made authentication table stakes.
LinkedIn got stricter, too
Here’s the part you also have to consider, LinkedIn throttles connection requests on a weekly basis — most accounts land somewhere around 100 a week, or roughly 20–25 a day. But treat that as a working figure, not a hard number. LinkedIn doesn’t publish the threshold, changes it regularly, and sets it per account.
Which makes LinkedIn a precision channel, not a volume channel
Do the math at that rough figure. Call it 100 requests a week at a 30% acceptance rate — about 30 new conversations per week, per account. Whatever the exact ceiling is on your account, it’s a ceiling.
Most founders read that as a limitation. I’d encourage you to read it as a feature.
The cap is exactly why per-message reply rates stay high. The channel structurally punishes spray and pray, which means it rewards the thing you should be optimizing for anyway: precision. You can’t out-volume anyone on LinkedIn, but you can absolutely out-target them.
So the question stops being “how do I send more?” and becomes “how do I make sure the ones I send are the right ones?”
That’s lever one, and it’s where signal-based GTM earns its keep. Instead of building a static list that matches your ICP on paper, you trigger sequences off evidence that someone is in motion right now — a funding round, a relevant job posting, a new exec in the buying role, a competitor switch, someone viewing your profile or engaging with your content.
The test: every campaign should be traceable to a trigger you can name in the first line of the message. “Companies that match our ICP” is a starting universe. “Companies that match our ICP and just did X” is a campaign.
The LinkedIn reply data makes the case better than I can:
Messaging existing connections: 16.86%
Messaging attendees of a LinkedIn event you both attended: 14.21%
Messaging people who viewed your profile: 13.40%
Template-based cold campaigns: 8.62%
Every motion that starts from a signal beats the cold blast by five points or more. Even a small signal helps — adding a profile visit before your message nearly doubles reply rate on its own (4.88% → 11.87%).
When your weekly sends are capped in the double digits, spending them on people who already have an observable clue is the entire game.
The hard part isn’t agreeing with this — it’s catching the triggers while they’re still fresh. Most founders and sellers check for them manually, which means they find the funding round three weeks after it was announced.
Watching them automatically is exactly why we built ABM Intel: it monitors your target accounts for 23+ event types — funding, leadership changes, layoffs, M&A, hiring surges — and routes each one to the rep with the evidence and a recommended next action. Whatever you use to surface them, the point is that a signal you notice a month late is just a cold list again.
The benchmarks to target
Here’s where performance sits today, so you have something real to measure against.
Sources: Expandi (13.2M requests, May 2025–April 2026), Cleverly, SalesSo.
The most useful thing on that table isn’t the numbers. It’s the diagnostic underneath them:
A low acceptance rate is a targeting and sender profile problem.
A low reply rate is a message problem.
Fix them in that order.
One more thing that makes this fast to read: 63% of acceptances land within 24 hours and 88% within a week. You don’t need a month to know whether a batch is working. If you’re under 25% acceptance after a week, stop sending and fix the list.
The rest of this piece runs in that order — profile, then message, then cadence, then tools.
Start with your profile
This is the single biggest gap between founders who make LinkedIn outbound work and founders who conclude it doesn’t.
Here’s what the data actually says about personalization, and it’s more useful than the version you’ve probably heard. Adding a note to a connection request barely moves acceptance — Belkins measured 26.42% with a note versus 26.37% without, across 20M+ outreach attempts. What it moves is what happens after: requests sent with a personalized message saw reply rates nearly double, 9.36% versus 5.44%.
So if the note isn’t what gets you in the door, what is? Largely, who you appear to be. C-level senders see 29.4% acceptance versus 26.3% for junior and entry-level senders. Prospects are still less likely to respond to a sales title. So if you’re a busy founder (who isn’t), share your login with your team and let them play magician behind the curtain — but use your identity for the outreach.
Which leads to the uncomfortable part. Your acceptance rate is downstream of your profile, and your profile is downstream of whether you’ve been visible at all. Someone who has seen your name in their feed twice is a different prospect than someone who hasn’t.
You can’t outwrite a profile that doesn’t hold up.
This is why founder brand and thought leadership is such a reliable GTM multivitamin — it isn’t a separate initiative from outbound. It’s the thing that makes outbound work. Your profile can boost your connection rates and your regular content continues to nurture that growing audience.
Founder Self-Check. Pull up your own profile on mobile, the way a prospect will see it. Honestly:
Does your headline say what you do for whom, or does it say your job title? CEO of [Startup no one knows] isn’t doing enough work.
Have you posted in the last two weeks?
Would a stranger understand your company from the first screen, without scrolling?
Does your photo look like a person you’d take a call with?
If you answered no to two or more, fix that before you spend a week’s worth of requests.
Write shorter than feels right
LinkedIn’s own data is unusually clear here.
InMails under 400 characters get a 22% higher response rate than average. Messages over 1,200 characters run 11% below average.
And here’s the opening: only 10% of all InMails are under 400 characters, while 46% are over 800. Being brief makes you stand out from roughly 90% of the inbox on format alone, before anyone reads a word.
Keep it to 50 words or less, curious and friendly. Ask, don’t assume. Lead with the trigger — the reason you’re reaching out now should be the first thing they read. It should sound like something you’d actually say out loud to someone at a conference. If your first message requires scrolling on a phone, it’s a pitch, not an opener.
Two follow-ups. Then stop.
Follow-ups behave differently on LinkedIn than in email, and this one surprised me.
Across 70,130+ campaigns, Expandi found the first follow-up produced 0.6% fewer responses than sending no follow-up at all. The second follow-up is where the lift is — 4.05% more responses. Third and beyond adds at most 1%.
So: plan for two follow-ups, save your best angle for the second one, and stop there. Note that this is the opposite of email, where 42% of all replies come from follow-ups and 3–5 of them is the recommended range. Don’t port your email cadence over.
Related: replies to connection notes have dropped from 3.5% to 2.2% in twelve months — a 37% decline. Notes can still help you get accepted. They just don’t start the conversation anymore. Write a short, specific note to earn the accept, and save your real effort for the message after they say yes.
The tool stack
Building your lead list
Same as 2024: Sales Navigator for filtering by network and firmographics, Apollo for contact data and intent. They’re complementary, and together they run under about $200/month. I go deeper in the outbound playbook.
The trap: tools that look like they automate LinkedIn
This is the one I want you to read carefully, because the product pages are genuinely misleading.
“I already pay for HubSpot and Apollo. Can’t I just sequence LinkedIn from there?”
Both will happily let you add a LinkedIn step to a sequence. So it looks solved. It isn’t — because that step doesn’t send anything. It creates a task telling you to go open LinkedIn and do it by hand.
Your sequence still runs. Your dashboard still shows LinkedIn touches. But a human being has to physically execute every single one, and the moment that person gets busy, the LinkedIn half of your cadence quietly stops while the email half keeps going.
To be clear, this isn’t a knock on either tool. They automate email end to end — send, sequence, track, log — and they do it well. The gap isn’t a roadmap decision or a technical limitation. LinkedIn simply hasn’t exposed an API and given them permission to do it.
I’ve asked my HubSpot reps about this more than once. Nothing I’ve heard suggests that barrier is lifting anytime soon. The day it does, this whole category of tool becomes unnecessary and I’ll happily tell you to cancel it.
Until then, you need something purpose-built.
I say that reluctantly, because my general position is to consolidate vendors aggressively. Every tool you add is another integration to maintain, another data silo, another line item, another login. Most startups I work with need fewer tools, not more. This is one of the few places where a separate tool is genuinely required rather than nice to have.
The tool I use and recommend
HeyReach (I used to recommend Meet Alfred but have found HeyReach to be far better).
Campaigns run whether or not your laptop is open, and it automates the full sequence: profile view, connection request, message, follow-ups, voice notes. Everything lands in one unified inbox which makes it easier for others to manage on your behalf. It starts at $79/month per LinkedIn sender, with a 14-day free trial that lets you connect up to three accounts without a credit card.
Two reasons I landed on it over the alternatives:
It integrates natively with HubSpot. If I’m going to break my own consolidation rule, the tool has to write back to the system of record. LinkedIn activity that never reaches your CRM isn’t measurable — which means you can’t tell which segments are working, and you can’t attribute a dollar of pipeline to the channel. It also syncs with your email sequencer, so LinkedIn and email run as one cadence instead of two disconnected motions.
It has an MCP server. Honestly, an MCP is a requirement for me nowadays - not just a nice to have. It means I can design and launch campaigns in plain language through an AI assistant — split lead lists, draft the sequence, tag replies by sentiment, push them to the CRM — instead of clicking through campaign builders. For a small team, that collapses a real amount of setup work.
Two things worth knowing if you’re early:
They run an early-stage program — a heavily discounted plan if you’re under $250K ARR with fewer than five people. Worth applying for.
Because LinkedIn throttles each account — not each user — the only way to genuinely scale this channel is multiple senders (you, a co-founder, your first AE). HeyReach is built around that — it charges per sender, not per user, so you can add teammates and VAs for free.
And the DIY version is completely legitimate: 20–25 connection requests a day, sent by hand, is a real strategy. It’s maybe 20 minutes. Founders regularly outperform automated campaigns doing exactly this, because the personalization is real.
One honest caveat
Every LinkedIn automation tool — HeyReach included — operates on platform tolerance, not platform permission. LinkedIn’s user agreement doesn’t sanction automated activity, and accounts can get restricted.
So use with caution. Ramp gradually rather than starting at the ceiling. Spread sends across business hours instead of firing a week’s worth on Monday morning. Treat withdrawn or ignored invitations as a signal to slow down, not push harder.
Be careful when automating the account that is your company’s most valuable brand asset. A restricted founder profile is recoverable (usually within 24-48 hrs in my experience), and easier to recover from than a burned sending domain.
Where to start
Good, better, best.
Good. Pick 100 accounts that match your ICP. Send 20 connection requests a day by hand, no tools. Short note. Real message after they accept. Two follow-ups, then stop.
Better. Build the list in Sales Navigator or Apollo. Send manually from the founder’s profile. Log every conversation in your CRM so you can see acceptance and reply rates by segment — you can’t improve what you can’t see.
Best. Trigger sequences off signals so you’re starting warm. Automate the sequencing across multiple senders. Run LinkedIn and email against the same list so a prospect who ignores one channel gets a second chance on the other.
Remember that outbound is still the lowest-quality pipeline you’ll build (relatively speaking). But run on LinkedIn, off a real signal, from a profile that holds up, it performs several times better than the 1% baseline — and that’s a gap worth closing while you build the other three categories.
You’ve got this!
📘 Get the full outbound guide
This article covers the LinkedIn half. The complete guide covers the rest: choosing your medium, building your list from signals, subject line research, cadence by segment, domain warm-up and deliverability, and the full tool comparison.
With love and gratitude -
If you want to learn more about working with me directly…
For B2B startups I serve as a Fractional GTM executive or advisor. Learn more about-
When you’re ready, let’s connect to discuss your specific growth goals and challenges.
Subscribe for weekly education, ideas, and frameworks
In this newsletter I share the exact tips, playbooks, and GTM multi-vitamins I’ve used to help 30+ B2B startups scale their revenue 150-590% YoY.
Thanks for reading GTM for Startups by Jess! This post is public so feel free to share it.