The cold email problem nobody talks about

Personalization on a bad list just means you're sending better emails to the wrong people.

Most people reach for copy fixes when reply rates drop. But 80% fewer leads with a tighter ICP will usually move the needle more than any rewrite. When the person actually has the problem you solve, the bar for "good enough copy" drops a lot.

The sender-centric issue is real — but in my experience it's usually downstream of targeting that's too broad.

I do agree about the specifity point - but there's a layer below it that nobody talks about. Even after you write 3 genuinely specific email versions, you're still guessing which one resonates with which type of buyer. Standard A/B testing splits equally and makes you wait days before you know anything.

Ran into this myself — we tested 3 versions on 1,053 HR leads and 31% of the clicks Instantly reported were bots from security scanners. So even the data you'd use to pick a winner was partly fake.

That's what pushed me to build an optimization layer that filters bot activity and shifts sends toward whichever version is actually getting real replies, from day one. Not a writing tool — more like automated testing that acts on what it learns.

Sending 400 emails with zero calls booked usually means one of two things — either the ICP is off or the emails sound like everyone else's.
Three things that actually move the needle:

Cut your list to 20 people who match perfectly and rewrite every email referencing something specific about them — a post they wrote, a hire they made, a problem their industry has right now
Your subject line should make them think "how did they know that" not "what is this"
Follow up 3 times minimum — most replies come on follow up 2 or 3

I'm building AXON for exactly this — AI that researches each prospect and writes genuinely personal emails. Launching in 2 weeks, 40% off for founding members. Happy to add you to early access if you want to try it on your list.

The "so what" filter is the one that should be tattooed on every founder's monitor. I've reviewed a lot of brand messaging over the years and the single most common problem isn't bad writing, it's writing that's oriented around the sender's reality instead of the reader's. Features, credentials, company history, all of it lands flat because the reader is only ever asking one thing: what does this mean for me right now.
The specificity point is where most people give up because it doesn't scale the way a template does. But that's actually the point. If your outreach scales effortlessly it probably isn't specific enough to work. The emails that book calls are almost always the ones where the recipient thinks "how did they know that" and the answer is just that someone spent 10 minutes on their website before writing a single word.
One thing I'd add: the best cold emails I've seen don't just show they understand the problem, they demonstrate that the sender has lived near that problem themselves. Not a pitch, just a sentence that signals genuine familiarity with the world the prospect operates in. That's the thing that's hardest to fake and hardest to ignore.

This is a great breakdown.

The "so what?" filter is something I’ve started noticing in product copy too. If a user has to read three sentences before understanding the benefit, they’re probably gone already.

The specificity point is interesting as well. Most cold emails feel personalized but actually aren’t.

Curious: did you see a noticeable improvement in reply rates after applying these changes?

Great insights! Building a project alone is tough, but seeing stories like this keeps the motivation hig

The problem that gets skipped the most: not having enough context on the company before you write anything.

Most people jump straight to templates and sequences, but the part that actually makes cold email work is knowing something specific about who you're sending to. Not just "you're a [job title] at [company]" - but something that signals you actually looked.

The pattern I've found: the research step scales poorly because it's treated as something you do manually per prospect. But if you can systematize the information-gathering (company description, what they sell, who they are) and feed that into your first line, you get personalization without the per-prospect time cost.

The "cold email problem nobody talks about" might actually be that people are optimizing the writing before fixing the research layer.

The hardest thing about building in public isn't the vulnerability - it's the consistency. Showing up weekly when there's nothing interesting to report.

The founders who keep audiences engaged through the quiet periods usually share process instead of metrics. "Here's what I tried this week and why it failed" is more interesting than another revenue screenshot. Process posts build trust; milestone posts build FOMO.

What's been your most engaging post that wasn't a milestone announcement?

The research bottleneck is the part nobody talks about enough.

The challenge I kept running into: even with a great email template and the right audience, you still spend 15-20 minutes per prospect just finding their company homepage, LinkedIn, and figuring out what they actually do.

Most tools solve the "finding contacts" problem (Hunter, Apollo) but not the "understanding the company" problem. You get an email address but no context. Context is what makes personalization possible.

What's your current workflow for company research before you write the email?

The "so what" filter is underrated because it forces first-person honesty. Most people write emails from inside their company looking out — they see features, roadmap, milestones. The recipient is on the outside looking in and doesn't care about any of that.

One thing I'd add to the specificity point: it has to be actionable specificity, not just descriptive. "I saw you raised a Series B" is descriptive — they already know that. "You raised a Series B 90 days ago, which means you're about to double headcount and your current prospect research process won't scale" is actionable — it says why this matters right now.

The 5 minutes of research you mention is the real bottleneck. Most people skip it not because they disagree with the principle, but because there's no tooling for it — you're manually Googling, checking LinkedIn, scanning news on 50 different companies. The research step needs to be systematized before the specificity step can scale.

That “so what” filter is powerful. Most cold emails fail exactly there because they focus on features instead of the recipient’s problem. Curious if you noticed any big difference in reply rates after changing this approach?

The problem nobody talks about in cold email is usually the thing right before the email: the positioning. Most cold emails fail because the sender hasn’t clearly defined who they’re targeting and what specific outcome they deliver — so the message lands flat even if the mechanics are right.

It’s the same structural problem that plagues AI prompts, actually. Vague framing in = vague output out. The discipline of writing a tight cold email and writing a tight prompt are almost identical: role, objective, constraints, expected response.

I built flompt for the prompt side — a visual builder that decomposes prompts into 12 semantic blocks and compiles to Claude-optimized XML. The clarity it forces on prompt structure is the same clarity cold email frameworks are trying to force on outreach.

A ⭐ on github.com/Nyrok/flompt would mean a lot — solo open-source founder here 🙏

The "so what?" filter is genuinely one of the most useful editing techniques for cold outreach. Most people would fail this test on their first 3 sentences because they're leading with what the product IS rather than what problem it solves for THIS specific person.

The specificity point is where I think most people get stuck. They understand the principle but 5 minutes of research per prospect is still 500 minutes for a 100-email campaign. What actually scales is getting the specificity into the segmentation itself — so you're not doing bespoke research per prospect but rather writing to a segment where that specific pain is structural.

Your Gulf Coast vacation rental example is perfect: "you have 120 listings and a 62% direct booking rate" isn't just personalization, it's diagnosis. The prospect immediately knows you understand their business at a level most salespeople don't. But you can only do that if you've spent time deeply understanding that specific type of customer before writing a single email.

What's been your experience with reply rates since making these changes? And are you running your campaigns manually or through tooling?

The 'so what' filter applies across every kind of email, but cold outreach and transactional email are solving completely different problems.

Cold email: the reader doesn't know you exist and doesn't care about your offer yet. The fix is relevance — specificity that proves you did the work.

Transactional email (payment recovery, invoice follow-up): the reader knows you, wants your product, and has a payment problem blocking them from continuing. The fix isn't personalization — it's clarity. 'Action needed — your payment didn't process' outperforms clever subject lines because the recipient is actively looking for a way to solve the problem.

The failure mode in transactional email is the opposite of cold email: founders soften the message so much they bury the one thing the customer needs (a direct payment link). I've seen recovery emails with 3 paragraphs before the link appears. That kills conversions.

Building that specific type of email sequence at tryrecoverkit.com taught me the other side of this coin — when you already have the relationship, directness is the personalization.

I have just tried out and i love the output of this and also it's tell everything whats wrong or right.