Cold Email Copy vs. Targeting: What Wins

What Caught My Eye on LinkedIn This Week

LinkedIn was loud this week. Lots of takes on cold email copy, deliverability, channel fit, and AI. Some of it was genuinely useful. Some of it was the usual noise dressed up as insight.

I went through the posts that got traction and picked the ones worth reacting to - the good, the incomplete, and the ones that need a reality check before you act on them.

Let's get into it.

The Copy Problem Nobody Wants to Admit

The 75-word guideline and the "flip sentences to be about them" advice - both correct. I've sent millions of cold emails and the pattern is consistent: the moment a sender starts talking about themselves, the reader's eyes glaze over and they hit delete. Every sentence that starts with "I" is a sentence the prospect doesn't care about.

The no-attachment point is also worth flagging. Attaching anything in a first cold email is one of the fastest ways to trigger spam filters. Keep attachments for follow-ups, after you've already gotten a reply. If you want templates that are already structured around the prospect, not the sender, grab these cold email templates - they're built with this exact principle baked in.

Context Is Everything - One Email Can't Serve Every Purpose

This is one of the more underrated points in the whole cold email conversation. People screenshot a "great cold email" and try to copy the format without asking what job that email was doing. A long personal narrative might work when you're selling yourself as a candidate. It will absolutely crater your reply rate when you're selling a SaaS product to a VP of Operations who gets 200 emails a day.

The rule of thumb I use: the more transactional the ask, the shorter the email. The more trust-dependent the ask (executive coaching, high-ticket consulting, niche services), the more context can help - but even then, "long" means 150 words, not 600.

If you see advice online that seems to contradict what you know, the first question is always: what vertical was this written for? A job seeker email and a B2B service email are completely different animals.

Channel Fit Is a Strategy Decision, Not a Preference

Correct, and worth repeating louder. The channel should follow the buyer - that line alone would fix a lot of broken outbound programs.

The most common version of this mistake I see: agency owners who hate cold calling decide to only do cold email, even when their target market (local contractors, small trades businesses, restaurants) almost never checks a business email inbox. They spend three months tweaking subject lines on campaigns that were dead before they started, because the channel was wrong from day one.

For B2B buyers - SaaS decision-makers, agency owners, consultants, mid-market executives - cold email works. These people live in their inboxes. But if you're chasing buyers who operate primarily through their phone and WhatsApp, you need to go where they actually are. No amount of A/B testing your subject line will fix a channel mismatch.

The Hyper-Specific Angle That Actually Books Meetings

The Teterboro airport example is a perfect illustration of what specificity actually means in cold email. It's not about adding a first name merge tag or referencing someone's LinkedIn bio. It's about narrowing your pitch to a version of your service that feels made for this exact person in this exact situation.

The staffing client example is the one I'd focus on - 40+ leads from a campaign built around one specific job role. That's the playbook. Most people write campaigns around broad categories of service. The teams generating real pipeline are writing campaigns around one micro-problem for one micro-segment.

I do agree with the caveat about overdoing it. The "guarantee you $50k" angle has been burned to the ground. Any email that leads with a massive revenue promise now reads as a scam before the prospect even finishes the sentence. Specificity in the offer, not the outcome claim, is where the work actually happens.

The Free Diagnostic CTA That Hit 90% Booking Rate

The numbers are impressive, and I think the mechanism is sound. Replacing a meeting request with a free diagnostic removes the commitment friction that kills most cold email CTAs. "Book 30 minutes with me" requires the prospect to give up time to a stranger. "Here's something free and useful" is a completely different ask.

I've seen this work with clients across multiple industries. The pattern: lead with a low-commitment deliverable - an audit, a quick analysis, a diagnostic call framed as research - and conversion from positive reply to booked meeting goes up significantly. The prospect who says yes to the free thing has already told you they have the problem and they're willing to engage. Closing from there is much easier than trying to close cold.

The caveat: this approach works best when your free diagnostic actually delivers something the prospect values immediately. If it's just a disguised sales call with a different label, people figure that out fast and it burns the approach for everyone else. Instantly is a solid tool for running this kind of campaign at scale if your infrastructure is set up correctly.

Deliverability Is the Language Your Campaign Speaks First

The analogy is a little soft but the underlying point is real. Deliverability is the foundation everything else sits on. The best-written, most hyper-personalized cold email in the world is worth nothing if it lands in spam. I've watched companies burn months of effort on copy optimization while their domain reputation was tanking and half their emails never reached an inbox.

The checklist before any campaign goes out: domain age and warm-up status, SPF/DKIM/DMARC configured, sending volume ramped correctly, list validated before import. Skipping any of these and then wondering why reply rates are low is like blaming the message when the letter never left the post office. If you need a validated list before you start, ScraperCity's email validator will clean your list before it does damage to your sender score.

The AI Training Data Problem Nobody Talks About

This is the most uncomfortable truth in the current AI-for-outreach conversation, and almost nobody addresses it directly.

Every LLM that generates cold email copy is trained on the internet's existing corpus of email content - which means it's learned from an enormous volume of bad outreach. The mediocre spray-and-pray emails, the feature-dump intros, the "I hope this finds you well" openers. The model doesn't know these emails had 0.3% reply rates. It just knows they exist and it patterns against them.

When you ask an AI to "write a cold email," you're statistically more likely to get something that mirrors average-to-poor output than something that mirrors what actually works. The solution isn't to avoid AI - it's to feed it better inputs. Your proven-performing email examples, your specific ICP details, your offer framed the way it actually converts. The AI is a drafting tool. If you want it to produce something useful, you have to bring the expertise. If you want prompts that actually produce decent cold email output, these Cold Email GPT prompts are built around the structures that convert, not what the average sender writes.

Links in Cold Emails: The Real Rule

The direction is right, the reasoning needs a little more precision. Links in cold emails hurt you in two ways: deliverability and intent signal. On the deliverability side, multiple links - especially to new or unrecognized domains - increase the chance your email gets flagged. On the intent side, a cold email stuffed with website links and product catalogue URLs signals that the sender wants the prospect to do research homework, not start a conversation.

The actual rule: first email, no links if you can avoid it. One link maximum if you genuinely need one - a Calendly for booking, a case study page if it's highly relevant. Never multiple links. Never your homepage. Never a product catalogue. The goal of the first email is a reply, not a product tour. Keep everything out of the email that isn't moving toward that single goal. For follow-up sequences, the rules shift slightly - check these cold email follow-up templates to see how links get introduced progressively across a sequence.

The 1,500 Email Experiment Worth Watching

I respect the commitment to actually testing it rather than debating it. That's the right instinct.

A few things I'd tell anyone about to run this kind of experiment before they hit send on email number one.

First: 50 emails a day is only useful if the emails are right. Volume without a working template is just faster failure. Before you scale up, test three to five completely different angles - different subject lines, different openers, different CTAs - and send 30 to 50 of each before drawing conclusions. You want to know which template works, not just whether cold email works.

Second: the follow-up is where most of the results actually come from. I've seen campaigns where the first email got almost nothing and the third follow-up had a 15% positive reply rate. If you're not building a three to five touch sequence, you're leaving most of the results on the table. The people who say cold email doesn't work usually stopped after one email and called it dead.

Third: track the data at the template level, not just the overall campaign level. You need to know which subject line gets opens, which opener gets replies, and which CTA gets bookings - separately. Aggregate numbers won't tell you what to fix.

I genuinely want to see the results of this one. The transparency commitment is the part that matters most.

The Bigger Pattern This Week

Reading through everything this week, the same underlying theme keeps surfacing: people are optimizing the wrong variable.

They're spending hours on subject lines when their domain is in spam. They're testing copy angles when they're selling to buyers who don't use email. They're adding links and attachments and product catalogues to emails that should have one sentence and one ask. They're using AI to generate copy without giving it the inputs that would actually make the output useful.

The order of operations matters. Get deliverability right first - your domain reputation, your list quality, your sending volume. Then get your targeting right - are these actually the buyers for this channel? Then get your offer right - is there one specific, low-commitment ask? Then worry about copy.

Most people start at copy and never get to the other three. That's why their results don't change no matter how many subject line tests they run.

In Cold Email Manifesto, the benchmark I set for a working campaign is a lead generator producing 200 clean, verified leads per day. Not 200 names - 200 contacts with first name, last name, title, company name, website, and verified email. The quality of that list is doing more work than any subject line ever will. If your list is junk, nothing else in the system can save the results.

The teams getting real pipeline right now are the ones who treated targeting as the primary lever, not the afterthought. If your list needs work before your next campaign, ScraperCity's B2B email database is worth a look - it's what I use to make sure the foundation is right before anything else goes out.

The one thing to take from this week: before you rewrite a single word of your cold email, check whether your emails are actually reaching inboxes, whether you're targeting buyers who use email as a primary channel, and whether your CTA is asking for the smallest possible commitment. Fix those three before you touch the copy. The copy problem is almost never the real problem.