1 week in: 3 things I got wrong about cold outreach (before sending a single email)

Shipped my product last week. Plan was: 100 cold emails to founders by Friday using a sequencer with a warmed-up domain. Pulled the brakes on day 3, before sending anything. Now I'm building a manual list of 20 names by hand, no tools. First batch goes out this week.

Here's what made me flip — three things I'd assumed were right, that turned out backwards.

1. Cold email tooling is a vanity step.

I spent two days reading about SPF, DKIM, DMARC, inbox warmup, sequencer setup, AI personalization. None of it mattered yet. The actual blocker: I didn't know who to email. I had a vague ICP ("small SaaS teams") and a list of company names pulled from random Twitter threads.

A sequencer with bad ICP is just scaled spam. Same domain rep cost, much wider blast radius. The tool stack would've let me waste 100 attempts before I figured out the ICP was wrong.

2. "Founder advice" in DMs is often lead-gen in disguise.

In the first 48 hours after launch, two commenters showed up with the same shape: abstract wisdom ("the decision sitting underneath the feedback"), zero specifics, no example, no question — closing with "drop your email and I'll send the tighter version."

That's not peer founder advice. That's a lead capture funnel using your launch as the trigger.

The tell: real peer founders drop one concrete heuristic in public ("I trust users who unprompted come back in week 2, not week 1"). Lead-gen coaches give abstract noun phrases and route to email. If they already have "the tighter version" pre-packaged, you're the lead.

3. Five minutes of manual research is the actual moat.

Now I'm doing this: open LinkedIn, find one specific recent post by my actual target (a Head of Data or founder), write an 80-word email that opens with a real reference to what they said. ~10 minutes per email.

The research step feels like distraction work. I'm betting that's exactly why most founders skip it — and why the inbox of a Head of Data is full of "AI-powered platform for SaaS teams" pitches that get archived in 2 seconds. One real, specific reference might just break through.

Will report back after the first batch lands.

For those of you who've made the manual → automated tipping point on cold outreach — what was the threshold for you? Volume? Reply-rate stability? Something else?

— Daniel, building saemi.at