Cold Email Offer First: What Actually Works Now

Video 1: Aaron Shepherd Builds a Cold Email Script Live (And Gets Most of It Right)

Aaron Shepherd runs GrowthFlare, has about 1,600 subscribers, and this video has just over 1,000 views. The title is bold: Watch me write a $250k cold email script in 20 minutes. I almost skipped it. I'm glad I didn't.

This is one of the more honest cold email videos I've seen from a smaller channel in a while. Aaron's core argument is something I've been saying for years: the offer is the 80/20 of cold email, not the copy. Most people spend 90% of their time agonizing over subject lines and first sentences. They never stop to ask whether their actual offer is something a stranger would respond to.

Here's the problem he identifies. Most businesses try to pitch their core service directly in a cold email. You're a stranger in someone's inbox asking them to commit to a $10,000 retainer based on one email. Aaron puts it plainly: that's never going to work. I've sent millions of cold emails. He's right.

The Offer Audit: What Aaron Actually Does in This Video

Rather than using his own client, Aaron challenges himself to work with a company he grabbed from ChatGPT on the spot: Disruptive Advertising, a full-service B2B growth agency. He looks at their existing offer: a risk-free guarantee, get results in 90 days or you don't pay, plus a free audit.

His critique of their offer is sharp and worth repeating. Two specific problems:

• No definition of results. What does results mean? A dollar amount? An ROAS target? If you can't define it, neither can your prospect. Vague guarantees aren't guarantees, they're just noise.
• The word audit is dead. I've been saying this for years. Nobody who doesn't know you wants an audit. To them it means getting on a 30-minute call with a stranger and wasting their time. The word audit has been run into the ground by every agency doing outbound.

He then runs his custom ChatGPT prompt to rebuild the offer, and Claude to write the actual email. The AI spits out three options. He walks through each one honestly, including calling out the ones that don't work. That intellectual honesty is rare in this space.

The offer he lands on: launch and validate three to five paid social creatives using internal benchmarks before the prospect commits to full management. They get real deliverables, not a vague promise. Hooks tested, winners identified, insights they keep even if they don't continue. That's a low-friction, tangible entry point.

Where I Agree Completely

The core insight here matches what I wrote in The Cold Email Manifesto. The email doesn't need to sell. It just needs to present the offer in the most casual, frictionless way possible. When your offer is strong enough, the copy is almost irrelevant. You're not building trust or establishing credibility in the email itself. You're just getting a response.

I've watched thousands of cold email scripts across the 14,000+ entrepreneurs I've worked with. The campaigns that fail almost always share one trait: they're pitching a service instead of an offer. There's a big difference. A service is what you do. An offer is what the prospect gets, at low risk, quickly, with a clear outcome.

This maps exactly to what Aaron is describing. When I worked with a magazine publisher through one of my consulting calls, their cold email was going out to 50 niche publishers pitching their full design and publishing service. Zero responses. The fix wasn't better copy. It was changing what they were asking prospects to do. Instead of pitching the whole engagement, we built an entry-point offer. Something specific, low commitment, and tangible. That's the same transformation Aaron is demonstrating live.

Where I'd Push Back

Aaron leans hard on AI for both the offer construction and the email copy. ChatGPT for the offer, Claude for the script. Nothing wrong with that as a starting point. But his Claude output has a real problem he actually spots himself and calls out: one of the generated lines says pinpoint which hooks actually convert without telling the prospect what they're actually getting. Is it a call? A document? An analysis? The prospect has no idea.

Vague deliverables kill response rates. If your offer is three to five creatives tested and validated, say exactly that. Don't dress it up in language that sounds like consulting speak. The cleaner and more literal the deliverable, the easier it is for someone to say yes.

He also leans into the Andromeda framing around Meta's algorithm update to add relevance to the outreach. I actually like this in theory. Timely, problem-aware context in a cold email is smart. But the execution matters. If everyone in the Meta advertising space starts leading with Andromeda references in their outreach over the next few months, it becomes noise fast. Relevance works until it's saturated.

What's Worth Implementing From This Video

One thing, and it's the most important thing: audit your offer before you touch your copy. Sit down and ask whether a complete stranger who has never heard of you would respond to what you're offering. Not your service. Your entry-point offer. If the answer is anything other than a clear yes, fix the offer first.

If you want templates built around strong entry-point offers, I have a set at killer cold email templates worth looking at before you write another word.

Aaron's channel is small. This video deserves more views. The framework is sound.

Video 2: Taylor Haren on Gmail's AI Inbox and What It Means for Cold Email

Taylor Haren runs a channel with about 2,900 subscribers. This video has 2,629 views and 85 likes, which is a strong engagement ratio for the size of the channel. The title: Cold Email is About to Change in 2026 (and nobody even realizes).

I watch a lot of content with titles like this. Most of them are thin. This one isn't.

The Core Claim: Gmail's AI Inbox Changes the Deliverability Game

Taylor's thesis is that Google's rollout of an AI-powered inbox, driven by Gemini, fundamentally shifts how cold emails get surfaced to users. Instead of a chronological inbox, Gmail moves to a brief-style, prioritized feed based on relationship signals. Who do you email frequently, who's in your contacts, what relationships can the AI infer from your message history.

The problem for cold email is obvious: cold emails have zero relationship signals by definition. Even if you have clean infrastructure, good domain reputation, and hit the primary inbox technically, the AI may still bury your email before the prospect ever sees it.

Taylor's framing of the shift: it used to be about landing in primary. Now it's about being seen within primary. That's a meaningful distinction and he's right to make it.

Is the Threat Real?

Here's where I'll give you my honest read. The underlying dynamic Taylor is describing is real. Email providers have been moving toward relationship-signal-based filtering for years. This isn't new as a concept. What's new is Google putting it front and center with a named AI product and a public rollout.

Taylor ran 9 million emails a month through a tool called Fixer AI, which he says does what Gmail is now doing natively. So he's seen this filtering behavior in practice. That's credible context.

He also makes a point I think a lot of people in the cold email world miss: enterprise and business workspace users will be slow to adopt this. IT teams are conservative. Companies that depend on their inbox for inbound leads aren't going to hand control to an AI filter overnight. There's a window here, and the size of that window matters enormously for anyone doing outbound at scale right now.

His read: use that window wisely. Don't assume you have forever.

The Tactical Advice He Gives

Taylor breaks down his response to this shift into a few categories:

• Infrastructure has to be bulletproof. SPF, DKIM, DMARC, warmed domains, rotated inboxes, residential proxies, daily blacklist monitoring. He keeps spam complaint rate below 0.3%, targeting closer to 0.1%. One complaint per thousand emails sent.
• Send smarter, not more. His team sends one email to the total addressable market every 60 days. They don't chase with five follow-ups in a week. They start with small volume subsets, validate the message, and only scale when the message is clearly working.
• Layer intent signals on top of volume. Company announcements, hiring activity, product launches. These trigger more timely outreach that's more likely to generate two-way conversation, which in turn builds the relationship signals the AI is looking for.
• Copy has to stop reading like a sales email. Short, punchy, direct. He caps it at 50 to 60 words. Write like you're messaging a friend, not writing marketing copy. Lead with a problem they actually have, not your solution.
• Multi-channel as a shortcut around relationship signals. If you can get someone to email you first, you bypass the problem entirely. LinkedIn conversations that move to email, content on YouTube or Reddit that generates inbound. He calls this inbound-led outbound and mentions they're building it as a service.

Where He's Exactly Right

The 50 to 60 word cap on email copy. I've been saying this for years and the data backs it up across thousands of campaigns I've reviewed. Long emails die. The moment your prospect feels like they're being sold to, their attention is gone. Short, direct, and framed around their problem is the only thing that works consistently.

His point about AI personalization is also worth highlighting. He references data from Eric Noowski confirming that slapping an AI-generated first line on the front of an email, the classic Hey Joe, I saw you commented about X on Twitter opener, doesn't land as human anymore. Prospects know. The bar for what reads as genuine has gone up, and fake personalization is actually worse than no personalization because it signals immediately that this is mass outreach dressed up as something else.

I've seen this in my own campaigns. Genuine first lines based on actual research still work. Templated AI personalization that follows an obvious pattern is getting tuned out. The difference is whether the line could only apply to that specific person or could apply to any of the 10,000 people on your list with a find-and-replace.

His multi-channel framing is also something worth taking seriously. If you're only relying on cold email, you're playing a harder game than you need to. Getting someone to initiate contact with you through content, LinkedIn, or any other channel fundamentally changes your standing in their inbox. You're no longer a stranger. For the AI inbox Taylor is describing, that distinction is going to matter more over time.

Where I'd Add to His Thinking

Taylor talks about sending one email to the TAM every 60 days as the core cadence. I think that's directionally right for protecting deliverability and avoiding spam complaints. But it puts enormous pressure on the quality of the single email you're sending. If your offer and your copy aren't dialed in before you go broad, you've just burned your 60-day window with a mediocre message.

The sequence has to be: nail the offer, validate the copy with a small list, then scale. Not the other way around. This is exactly the point Aaron Shepherd makes in the first video, and it's not a coincidence both videos land on the same conclusion from different angles.

On the infrastructure side, if you're serious about deliverability in this environment, you need to be checking your domains daily and replacing anything that gets flagged fast. The tools I use for this: Smartlead for managing inbox rotation at scale, and Instantly for warming and monitoring. Both have gotten meaningfully better at surfacing deliverability issues before they compound into serious problems.

For building targeted lists with the intent signals Taylor mentions, hiring activity, company announcements, recent product launches, I run those through ScraperCity. Being able to segment a list by what's actually happening at a company right now is one of the few things that still creates genuine relevance in a cold email at scale.

What's Worth Implementing From This Video

Two things immediately actionable:

First, check your follow-up cadence. If you're sending more than two follow-ups in a sequence and they're going out within a week of each other, you are actively training spam filters to treat your domain as a threat. Pull it back. Taylor's 60-day re-contact window sounds conservative until you realize it's what's keeping his deliverability clean at nine-figure email volumes.

Second, start building inbound signals. Even something as simple as being active on LinkedIn and driving replies back to your email creates relationship signals that offset the cold outreach problem Taylor is describing. It's not an overnight fix, but it compounds over time in a way that pure cold volume never does.

For copy, cap your emails at 60 words. If you can't say your offer clearly in 60 words, the offer isn't clear enough yet. Go back to the offer before you touch the copy. Check out new email scripts for examples of what short and direct actually looks like in practice.

The Pattern Across Both Videos

I watch a lot of cold email content. These two stood out this week for a specific reason: they're both pointing at the same root problem from completely different angles.

Aaron is building a cold email script and starts with the offer because the offer is what determines whether anyone responds. Taylor is preparing for an AI inbox that will filter by relevance signals, and concludes that if your offer is something people actually want, the AI will let you through. If it isn't, no amount of technical infrastructure saves you.

The throughline is this: offer quality is the only durable advantage in cold email right now. Copy can be tested and optimized. Infrastructure can be managed with the right tools. But if your offer is vague, high-commitment, and designed for your convenience rather than your prospect's, none of the tactical stuff matters. You're building a faster car on a road that leads nowhere.

This is the same thing I've seen in consulting calls with hundreds of entrepreneurs. The ones struggling with cold email almost always have an offer problem, not a copy problem. They're pitching their service instead of creating a low-friction entry point. They're asking for too much from a stranger who has no reason to trust them yet.

Fix the offer. Then fix the copy. Then scale the infrastructure. In that order. Every time.

For subject lines that work once your offer is dialed in, I put together a breakdown at cold email subject lines that covers what's actually moving the needle right now. And if you want to look at full sequence structures, cold email follow-up templates has the frameworks I use across different industries.

This Week's Single Actionable Takeaway

Write down your current cold email offer in one sentence. Now ask: would a stranger who has never heard of you respond to this with zero risk on their part? If the answer involves them committing time, money, or access to their backend before they trust you, the offer isn't ready. Rebuild it as a specific, tangible deliverable they can say yes to without feeling like they're committing to anything. Get that right first. Everything else follows.