We spent $5,000 a month on LinkedIn ads for a client and generated zero leads.
On purpose.
Most B2B companies treat LinkedIn ads as a lead-gen channel. That's the mistake. Your ICP is there, but the second you gate them behind a lead form, sales calls the "leads" and hears nothing back.
LinkedIn ads are a warm-up channel, not a lead-gen one. Here's the framework that got this client a 20-25% lift in positive replies 👇
@saguppa Point 3 is quietly the whole thing. Wiring the MCP and picking the data source are learnable in a weekend now. Reading a reply and knowing in one line whether it's worth a human is the part that doesn't get commoditised. The tooling got easy, so judgment became the actual moat.
The line that should reset how people think: optimise it like SEO, measure it like Facebook. That's why teams underrate AI search: they're looking for last-click traffic from an assisted, view-through channel. Gamma only saw the real number when they started asking, "How did you find us?" Most brands never ask, so they conclude it isn't working
The 44% from the first third stat should change how everyone writes. Most B2B content still opens with a warm-up paragraph before the actual answer. That warm-up is exactly what retrieval skips. The fix is uncomfortable but simple: answer in the first line, then add the context. Write the H2 as the question; the sentence under it as the answer.
We run HeyReach, so this one genuinely lands. Describing a campaign in chat and launching it kills the part everyone hates: the setup. What I like is that it removes the busywork without pretending to remove the judgment. You still decide who to reach and what earns the reply. That's the useful kind of AI.
@zackkanter Feeling a version of this from the non-dev side. We're building AI-native ops at the agency and the bottleneck stopped being ideas months ago. It's capacity and cost to actually run them at the speed the work now expects. Demand for "just do it now" completely outran supply.
The buyer-quality point matters most. Someone who booked after 3 weeks of your content vs someone squeezed out on the 7th follow-up churn completely differently. One arrived pre-sold; the other arrived sceptical. That said, cold isn't dead; spray-and-pray just is. Signal-based outbound to people actually showing intent still pulls the pre-sold kind. The real split isn't ads vs cold; it's inbound-quality intent vs interruption.
Signal-based wins because it fixes timing, not just targeting. The catch people miss: a signal is only worth the sentence it lets you write. Fire the same generic email off a pricing-page visit and you're back to cold with extra steps. The reply comes from referencing the signal, not just having it.
The channels aren't the problem. All three are broadcast, and your first 10 users don't come from broadcast; they come from you going one-to-one. Find the 50 people who already have the problem you solve in the communities and comment sections where they complain about it, and talk to them directly. No posting. Just conversations. The first 10 users are hand-to-hand, not a channel.
Reddit isn’t working for me.
TikTok isn’t working for me.
Instagram isn’t working for me.
Where can I find my first 10 paying users without spending any money on marketing?
We basically run this setup in-house- GSC and GA4- in one place, with Claude on top. The connection and the live dashboard are the easy 80% now. The 20% that matters is knowing which page-2 keyword is actually worth chasing and whether the fix moves a conversion, not just a ranking. The tool surfaces the wins. A human still decides which ones are real.
@paolo_scales Half of these come back to one thing: specific and auditable beats broad and safe. Vague claims feel safer to write and convert worse every time.
This is the right model. The phrase doing all the work is "feed each other." Most teams run four channels in four silos and call it multi-channel. The wiring is the whole thing, whether it's Clay or your own signal layer. Channels that don't talk to each other are just four half-budgets.
pen ChatGPT and ask what your buyer asks before they find you.
"Best tool for X." "Alternatives to [competitor]."
Does your name come up?
For most B2B companies, it doesn't. And that's quietly the most expensive gap in marketing right now. https://t.co/FOHmKuFJuU
@LoganTGott Notice almost every line here runs on one thing you already own: your sales calls. Objections, case study numbers, the way you actually talk. The founders who struggle with this list are the ones treating it as 17 new tasks instead of one asset repackaged 17 ways.
@DavidGQuaid EMDs mattering again in an AEO world is a genuinely interesting shift. The domain's the cheap part, though; the citable content behind it is where the real cost hides.
@alishaz1221 Hey! Ishan here, building GrowthSpree. We run growth for B2B SaaS (paid, ABM, and getting brands cited by AI). Always glad to meet people actually shipping. Let's connect.
@paolo_scales The number that quietly carries this whole model is 5,000 impressions per post, minimum, twice a week, forever. That's the actual hard part. Everything downstream is conversion tuning. Most people can't reliably clear that first line, which is why the funnel never fills.