Hey 👋
I was talking to a founder last week who still starts the week by exporting a cold list.
What struck me wasn’t the hustle. The hustle is fine.
The list is the leak.
Most of the replies that actually turn into meetings don’t come from a spreadsheet somebody bought. They come from people who already raised a hand - followed a competitor, liked a post in your category, changed jobs, showed up in a hiring signal.
Same channel. Same kind of message. Wildly different reply rates: 15-47% from those warm sources versus 1-3% cold.
Here’s what that actually looks like - eight places to check before you write a single LinkedIn message.
TL;DR
Warm outreach from raised-hand sources replies at ~15–47%; cold lists sit at ~1–3% - same channel, often the same message shape
LinkedIn’s practical cap is still ~100 connection requests / week - volume is no longer a strategy; list quality is
Signal-triggered lists see 40–60% accept vs 10–20% on cold titles (2026 outreach benchmarks)
H2 data: AI-personalized ONLY copy underperformed human templates by ~12% on acceptance - fake research is worse than a tight human line
Job-change is still a real budget reset - generic “congrats on the new role” is now the spam version of that play
Eight checks before send: competitors, adjacent tools, outbound job posts, job-changers in your network, specific creator magnets, ICP cut, recent activity, active-on-LinkedIn
Practical: delete low fits (and medium, if you’re aggressive) before the campaign launches - send volume is finite
Weekly finds
AI-personalized LinkedIn copy lost to human templates - by 12%. Expandi looked at 13M+ connection requests. Same accounts. The robot wrote faster. The human still got accepted. “Hyper-personalization” that smells like a scrape is just a new template.
Signal lists: 40–60% accept. Cold lists: 10–20%. You still get ~100 invites a week. Spend them on people who already leaned in, or spend them on job titles. LinkedIn does not care which you prefer.
Job-change still works. “Congrats!” does not. New seat, new budget, new problems. Your first line matching every other vendor’s first line is how you join the pile.
LinkedIn’s 2026 automation crackdown is a rate limit with teeth. ~100 requests a week, vendor-level bans, behavioral scoring. The volume bots got detention. Relevance is the new quota.
Anthropic backers talking $2T IPO after SpaceX. Claude Code is paying the bills. Distribution inside a workflow beats another demo. File under: same lesson as outbound, bigger numbers.
THE ISSUE
8 Places Before One Outbound Message
You’ll like this one if you run outbound.
Before we write a single LinkedIn message, we check the same eight places. Every time. It’s not exotic. It’s a habit almost nobody has actually built - because a cold list is faster to open than a warm one is to assemble.
Most of the good replies don’t come from a file somebody sold you. They come from people who already raised a hand somewhere. Followed a competitor. Liked a post next door to your category. Changed jobs. Showed up in a hiring signal.
Here’s exactly where we look.
1. People following your direct competitors
If someone follows the tools in your category, they’ve already decided the problem is worth solving. That’s the most obvious pool - and most teams stop here.
Don’t stop at “follows the company page.” Pull the people who viewed your profile, engaged with your posts, followed you, followed the company page, and engaged with competitors. One warm list. Not eight tabs.
Practical: If you only have time for one source this week, start here. It’s the shortest path from “they know the category” to “they might talk.”
2. Engagement on adjacent tools - not just direct competitors
We almost missed this one.
Someone on the team was watching likes and comments on a cold-email tool - not a direct competitor - and said the quiet part out loud: everyone engaging with that stuff cares about outbound somehow. Different product. Different channel. Same underlying intent.
Treat adjacent-tool engagement like competitor engagement. If they hang out where your problem is discussed, they’re in the pool.
3. Job postings for outbound roles (read as budget)
A company doesn’t post for an SDR or a BDR without money behind it.
One standing search: companies with open outbound roles, selling B2B, in your sweet-spot size. That query has surfaced some of our best-fit accounts before they showed up anywhere else - because hiring is a funded signal, not a vibe.
Hiring ≠ a named buyer in seat. The posting tells you budget exists. You still have to find the human who owns it.
4. Anyone in your real network who just changed jobs
A new role resets budget, authority, and problems all at once.
This isn’t a product feature. It’s a rule worth stealing. The moment someone you already know moves companies, a short “congrats - let me know if I can help” lands as a relationship continuing, not a cold pitch starting.
Skip the template congrats flood. One specific line about the new seat. Offer help. Don’t ask for 30 minutes in sentence two.
5. Engagement on one specific creator - not “everyone who liked anything”
We don’t scrape the whole internet.
Pick magnets whose audience actually overlaps yours - creators your buyers already read - and watch who engages with them. One creator in your niche, scraped daily, can feed a near-perfect list if you stay consistent.
General “engagement” is noise. Named magnets are a filter.
6. Score against ICP before anything gets sent
None of the above is a real list until you delete the people who don’t fit.
Rule before a campaign launches: cut every low fit. If you want to be aggressive, cut the medium fits too. Your send volume is finite. LinkedIn already decided that for you (~100 invites a week, and they mean it).
Spend it on people worth spending it on.
Practical: If you’re embarrassed to send the message to a name on the list, they shouldn’t be on the list.
7. Recent activity beats how good a company looks on paper
One of the best recent conversations started because someone was active on a competitor signal - not because their company looked perfect in a spreadsheet.
By the time we got on the call, the thread already made sense. They hadn’t found us. We’d found them while the interest was still warm.
Paper fit without motion is a nice logo. Motion without a perfect firmographic is often the deal.
8. Filter out anyone who isn’t actually active on LinkedIn
Before a message goes out, check whether they’ve shown up recently - posted, engaged, a real network behind the title.
A dormant profile with a perfect job title still won’t see what you send. This filter alone saves more wasted sends than almost anything else on this list.
Dead profiles eat your weekly cap. Live ones can actually reply.
Why the reply rates look “unfair”
Warm sources: 15–47%. Cold: 1–3%. Same channel. Often the same message shape.
The only thing that changed is whether the person had already raised a hand - somewhere - before you messaged them.
You don’t need eight tools. You need eight checks, then one honest first line.
Practical: This week, pick two sources (competitors + one magnet). Cut everyone who fails ICP or looks dormant. Send only that list. Compare replies to last week’s cold export. The spreadsheet usually shuts up.
