AI Tips

The 'Saw Your Post' Play That Hits 75% Acceptance

Why messaging the people creating content beats every other form of outbound we've tested

Tim Burnham, Founder & CEO

Tim Burnham

Founder & CEO

April 17, 2026

Most LinkedIn outbound starts with a question: "Who should I message?"

The honest answer for most teams is some flavor of "people whose job titles match our ICP." So they buy a list, fire up Sales Navigator, and start sending connection requests into a void. Acceptance rates hover around 15–25% on a good day. Reply rates are worse.

There's a much better question to ask: "Who in my ICP is creating content right now?"

Because the people writing posts on LinkedIn are doing two things at once. They're telling you what they care about, and they're proving they're active on the platform. That combination — high intent signals plus a guaranteed audience — is the foundation for one of the highest-performing outbound plays we've ever run.

One of our clients is currently sitting at a 75% LinkedIn connection acceptance rate with this exact approach. Here's how it works.

The Core Insight

Almost every "personalized" outbound message you've ever sent is actually about you. You researched the prospect, you found a hook, you wrote a clever line, you pitched.

The "Saw Your Post" play flips that completely. The entire opening message is about them — specifically, the post they just wrote.

"Hey [Name], saw your post about [Topic] — really resonated with the part about [specific thing]. Would love to connect."

That's it. No pitch. No research dump. No clever angle. Just a real human reaction to something they put out into the world an hour ago.

We've tested both. Heavily researched messages — the kind where you dig into someone's background, find a podcast they were on, reference their last funding round — underperform "Saw your post" messages. Why? Because researched personalization still feels like it's setting up a pitch. A genuine reaction to a fresh post feels like a peer reaching out. People can tell the difference instantly.

How the System Works

Step 1 — Define Your ICP by Topic, Not Title

Job titles are a blunt instrument. Topics are a scalpel.

Instead of "founders and CEOs," go a layer deeper: founders and CEOs who post about burnout, productivity, personal development, or company culture. Now you're not just targeting a role — you're targeting a mindset. Someone posting about burnout right now is in a completely different headspace than someone posting about Series B fundraising.

Pick 3–5 topics that align with what you sell, and build your search around those.

Step 2 — Scrape Posts From Your ICP

Use Apify (or a similar LinkedIn scraping tool) to pull recent posts that match your topic filters from people who match your ICP. You can layer in firmographic filters here too: company size, geography, industry, follower count.

Pro tip: targeting high-follower CEOs and founders specifically tends to produce better acceptance rates. People with bigger audiences are more comfortable accepting connection requests from strangers because their inbox is already noisy — one more thoughtful message stands out instead of feeling like an intrusion.

Step 3 — Use Engagement as a Quality Dial, Not a Brake

Here's the lever almost nobody uses: filter posts by combined likes + comments.

Set a threshold (we typically start at 5+ engagements) to ensure you're only messaging people whose posts are actually getting traction. A post with zero engagement might be a dud or a personal note. A post with real engagement is one the author cares about — which makes your "saw your post" message land harder.

But here's the philosophy I've landed on after running this play across a lot of clients: don't use this filter to throttle. Use it to tune.

My favorite thing to do is push a client right up to the edge of capacity. Crank the volume, lower the threshold, layer in adjacent topics, expand the geography — keep amping it up until they tell me their team can't physically reply to one more conversation. That's the right operating point. Most teams are running at 20% of what they could actually handle, and they don't realize it until you flood the top of the funnel and watch what their pipeline does over the next 60 days.

So: start at 5+ engagements, then experiment upward in volume, not downward. Lower the threshold. Add more ICP topics. Bring in a second persona. The only reason to raise the threshold is if message quality drops — not if the inbox feels busy. Busy is the goal.

Step 4 — Generate Personalized Connection Requests

For each qualified post, an AI step generates a short, personalized connection request that references something specific from the post. Not the headline — a specific line, a specific argument, a specific anecdote. The kind of thing only a human who actually read it would catch.

This is where AI earns its keep. Manually personalizing 200 connection requests a day is impossible. Generating them with a carefully tuned prompt is trivial.

Step 5 — Send Through Salesforge as Part of a Multi-Channel Sequence

The connection request is just the first touch. From there, Salesforge runs the full sequence:

  1. LinkedIn connection request referencing the post
  2. LinkedIn follow-up message if accepted but no reply
  3. Email #1 (plain text — see below)
  4. Email #2 as a final touch

A few important details we've learned the hard way:

  • Plain text emails outperform formatted ones. Every time. No fancy HTML, no logos, no signatures with five social icons. Just text, like a real person typed it.
  • Multi-channel sequences beat single-channel. LinkedIn alone is good. LinkedIn + email is significantly better.
  • Prioritize speed over volume. Configure your tools to surface real-time replies (within 24 hours) rather than maxing out daily send limits. The fastest response wins the conversation.

LinkedIn caps social actions (likes, comments, follows, connection requests) on a rolling basis, and your outreach tool will have its own monthly send credits on top of that. If you're running this play at the volume I recommend, you will hit the ceiling — and that's actually a good signal. It means you're pushing the channel for everything it can give you. Use the forced pause to catch up on replies and work the conversations you've already started, then go right back to redlining it.

What You Can Expect

Real numbers from a current client running this exact system:

  • ~5,000 founder leads identified in Clay
  • ~1,600 emails enriched via Trykit (about $25 in credits)
  • 75% LinkedIn connection acceptance rate on the highest-performing audience segments
  • 20–25% baseline acceptance across broader segments
  • 7.2% reply rate on accepted connections
  • 133 engaged conversations in active workflow at any given time

For comparison, "good" cold outbound lands around 1–2% reply rates. This system is doing 5–10x that on the LinkedIn side.

Why Email Underperforms LinkedIn (And What To Do About It)

We see this consistently: when you run LinkedIn and email side by side, LinkedIn wins by a wide margin. Cold email-only campaigns typically land around 0.5–1% reply rates. The LinkedIn arm of the same campaign — using the "saw your post" play — routinely produces 5–10x that.

The lesson isn't "stop doing email." It's "use email as the second touch, not the first." Email works much better when it follows a LinkedIn interaction — even one as small as a connection request being accepted. The prospect already recognizes your name. The email doesn't have to do all the work alone.

Our typical playbook:

  1. Start in LinkedIn. Connection request + post reference.
  2. Test new messaging in smaller segments first. Pilot in a smaller geography or sub-vertical before rolling out to your biggest one. If something's broken, you find out cheap.
  3. Split test ruthlessly. Salesforge and most modern tools support A/B on subject lines, opening lines, CTAs. Use it.
  4. Layer email in once LinkedIn is humming. Now you've got two touchpoints reinforcing each other.

Reply Handling: Manual Now, AI Later

This is the part nobody talks about: outbound is only as good as your follow-through on replies.

Salesforge has AI auto-reply built in. We don't always turn it on right away. For most clients, we start with manual responses — handled in batches of 10–15 at a time, with AI assistance for drafting — until volume becomes genuinely unmanageable. Then we layer in automation.

Why manual first? Because the early replies teach you what's working. The exact phrases people use to push back, the questions they ask, the objections they raise — that's your messaging optimization data. Automate too early and you lose that signal.

The Integration Gap Nobody Warns You About

Here's the honest part: none of these tools talk to each other cleanly out of the box.

LinkedIn replies live in Salesforge. Email replies live in Salesforge. But your CRM (we see this constantly with GoHighLevel and HubSpot) doesn't get a clean view of either without help. The result is two separate lead streams that have to be manually coordinated, and reps working leads in spreadsheets because the automation has gaps.

The fix is usually a small Make.com or n8n workflow that captures replies from Salesforge and pushes activity events into your CRM. It's not glamorous, but it's the difference between a system that scales and a system that requires a full-time human just to keep the data clean.

The biggest mistake we see: teams turn on a high-volume outbound system before they've built the integration plumbing. Then they drown in disconnected data. Build the Make/n8n glue between Salesforge and your CRM first, then pour fuel on the campaign. Your reps will thank you.

The Bottom Line

The "Saw Your Post" play works because it inverts the usual outbound dynamic. Instead of interrupting people, you're reacting to something they chose to put into the world. Instead of guessing what they care about, you're responding to a topic they just publicly endorsed.

When you combine that with the right ICP filter, the right engagement threshold, and a multi-channel sequence with clean reply handling, you get acceptance rates that look like typos.

75% isn't a fluke. It's what happens when every part of the funnel — targeting, message, channel, timing — is pulling in the same direction.

This entire system — Apify scraping, Clay enrichment, Salesforge multi-channel sequences, the AI personalization layer, and the Make/n8n glue that keeps your CRM in sync — is exactly what we set up for clients every week. If you'd rather skip the build and have it running by next Monday, book a free 30-minute call and we'll scope it out together. No pitch, no pressure — just a real conversation about what's possible.

Share

Let's Talk

We'll hop on a call, hear what's broken, and shamelessly pitch ourselves as the fix (of course). Or... if we're not the right fit, we'll tell you.