We Roasted 100 Cold Emails — Here's What They All Got Wrong

We didn't set out to study 100 cold emails. We just kept roasting them, and after a while, the same disasters kept showing up in the same order.

Sales reps, founders, freelancers, recruiters — different industries, different targets, different tools. But the same weak opener. The same buzzword soup in the middle. The same passive, timid close at the end. It started to feel less like individual bad luck and more like a shared playbook for getting ignored.

So we pulled 100 recent roasts through RoastMyEmail, stripped out anything identifiable, and looked at what the AI flagged most. The results were bleak — and extremely useful.

Here's what the data actually looks like.

The numbers first

Across 100 cold emails roasted:

• 94 out of 100 opened with a line that could have been sent to anyone
• 87 out of 100 had a value proposition buried past the third sentence
• 76 out of 100 used at least one phrase our AI flags as "buzzword bingo" — leverage, synergy, cutting-edge, scalable solutions, move the needle
• 71 out of 100 asked for a meeting before establishing any reason the recipient should care
• 68 out of 100 had a subject line under five words that was either too vague or too clever to earn a click
• Average cold email score: 31 out of 100

Thirty-one. The average email in our sample would score in the savage tier — which means our AI would spend most of its feedback asking who hurt you before offering a rewrite.

The good news: fixing even the top three mistakes on this list would push most of these emails from the 30s into the 60s. That's the difference between automatic delete and an actual reply.

Mistake #1: The opener that announces a bulk send

Showed up in: 94 out of 100 emails

"I hope this email finds you well."

"I came across your profile and wanted to reach out."

"I noticed your company is doing great things in the [industry] space."

These lines are not neutral. They are actively harmful. Every one of them signals to the reader — before they've finished the first sentence — that this email was not written for them specifically. It was written for a list they happen to be on.

The problem isn't politeness. The problem is genericness. An opener that could be copy-pasted to 5,000 people proves, in real time, that it was.

What to do instead: Open with something specific — a recent post they published, a hire they made, a problem you noticed, a result that's relevant to their situation. One sentence of genuine observation beats three paragraphs of polished nothing.

Mistake #2: Leading with your company, not their problem

Showed up in: 89 out of 100 emails

The second sentence of most cold emails in our sample read something like: "At [Company], we help businesses like yours..."

Here's what that sentence communicates to the reader: I have not thought about you at all. I am now going to talk about myself until you stop reading.

Cold email is not a brochure. Nobody opens their inbox hoping to learn about a company they've never heard of. They open it hoping something in there is relevant to a problem they're actually sitting with.

The switch is simple but requires a mindset shift. Stop leading with what you do. Start leading with what they're dealing with.

One of these earns a second sentence. The other earns the archive button.

Mistake #3: Buzzword soup in the value proposition

Showed up in: 76 out of 100 emails

Our AI flags certain phrases almost every time: cutting-edge, best-in-class, synergy, leverage, scalable, move the needle, unlock growth, streamline workflows, drive ROI.

These words have been in cold emails for so long they have lost all meaning. They no longer communicate anything. They just take up space where a real claim could live.

The test is simple: can you replace the buzzword with a specific number, outcome, or mechanism? If yes, do that. If you can't — if there's no concrete thing behind the word — that's a bigger problem than the language.

One of those sentences is specific enough to be believed or disbelieved. The other one is vapor.

Mistake #4: Fake personalization

Showed up in: 73 out of 100 emails

Surface personalization — a first name, a company name, an industry mention — is not personalization anymore. It's mail merge. Everyone's seen it. Nobody's impressed.

The emails that scored worst in our sample were the ones that appeared personalized but weren't. Things like: "I saw you're in the SaaS space" or "I noticed [Company] is growing fast" — lines that could be generated by pulling from a spreadsheet column.

Genuine personalization references something that required a human to actually look. A specific LinkedIn post. A piece of content they published. A job they recently posted. A customer they just won. Something that proves you spent two minutes on their world before writing.

The irony: real personalization doesn't need to be long. One sentence that proves you looked beats five sentences of generic flattery.

Mistake #5: The buried value proposition

Showed up in: 87 out of 100 emails

We found the actual value proposition — the thing the sender does and why it matters — past the third sentence in 87% of emails. In many cases, it didn't show up until the fourth or fifth paragraph.

By then, the reader is gone.

A busy VP, founder, or decision-maker is not going to read to the end to find out what you're offering. If your value isn't clear by sentence two, your email has already failed.

This usually happens because the sender is warming up to their pitch — trying to build context, establish rapport, earn the right to make the ask. It's a reasonable instinct. It's also the wrong one for cold email.

The fix: write your clearest, most specific value statement. Then move it to the top. Everything else exists to support it.

Mistake #6: Asking for too much, too soon

Showed up in: 71 out of 100 emails

"Would you be open to a 30-minute call next Tuesday to explore how we could work together?"

This is the cold email equivalent of proposing on a first date. You haven't earned a 30-minute meeting. You've sent one email. The ask should match the trust level — and at this stage, the trust level is zero.

The most effective CTAs in the emails that scored well asked for something small: a one-line reply, a yes/no, a reaction, a referral. Not a calendar hold.

One of those CTAs respects the reader's time and signals confidence. The other one sounds like every other cold email in their inbox.

Mistake #7: The multiple CTA problem

Showed up in: 58 out of 100 emails

"Check out our case study, book a call, or reply to this email if you have questions!"

Three options is no option. When a reader has to decide between multiple actions, the easiest decision is to take none of them. Decision fatigue is real, and cold email is not the place to test it.

Every cold email should have exactly one clear next step. Not two. Not "whichever works best for you." One.

Pick the lowest-friction action that moves the conversation forward, and ask for only that.

Mistake #8: The passive close

Showed up in: 66 out of 100 emails

"Let me know if you're interested." "Feel free to reach out if this resonates." "Happy to chat if the timing is right."

These closes put the entire burden of continuation on the reader. They communicate low confidence. They make the sender sound like they expect to be ignored — which, with lines like these, they usually are.

A strong close is direct without being aggressive. It tells the reader exactly what you want them to do and makes that thing feel easy.

Mistake #9: The signature longer than the email

Showed up in: 44 out of 100 emails

CEO. Founder. Award-winning. Forbes 30 Under 30. Certified partner. Five social media icons. A headshot. A legal disclaimer.

We've seen cold emails where the signature took up more vertical space than the actual message. This does not make you look important. It makes you look like you're compensating for a weak pitch.

Keep the signature to: name, title, company, one link. That's it. Let the email do the work.

Mistake #10: Subject lines that commit to nothing

Showed up in: 68 out of 100 emails

"Quick question" "Checking in" "Idea for [Company]" "Introduction"

These subject lines are technically inoffensive and practically invisible. They don't tell the reader why they should open the email. They don't create curiosity. They don't signal relevance. They just exist.

The best subject lines in our sample either referenced something specific to the recipient or made a clear, honest promise about what was inside. Both approaches outperform the vague placeholder every time.

If your subject line could apply to 10,000 people without changing a word, it needs to be rewritten.

What the top-scoring emails had in common

A small number of emails in our sample scored above 70. They were different industries, different offers, different lengths. But they shared a few consistent traits:

• They got specific fast. The reader knew what was being offered and why it was relevant within the first two sentences.
• They proved something. A number, a result, a named client, a concrete outcome. Not "we help companies grow" but "we helped a 12-person team book 40% more demos without adding headcount."
• They asked for something small. A yes or no. A two-line reply. Not a calendar invite, not a 45-minute discovery call.
• They sounded like a person wrote them. No buzzwords. No corporate filler. No "I hope this email finds you well." Just a human observing something relevant and making a clear, specific point about it.

None of this is complicated. All of it takes actual effort — which is exactly why most cold emails don't do it.

How to know where your email actually stands

Reading this list and nodding is easy. Knowing which of these mistakes are in your draft is harder.

That's what RoastMyEmail is for. Paste your cold email — a sales sequence, a LinkedIn DM, a recruiter message, an investor intro — and get a scored breakdown of exactly what's working and what isn't. Not vague advice. Line-by-line feedback, a scorecard across clarity, personalization, hook strength, and spam risk, and a rewrite direction you can actually use.

The average email in our sample scored 31. Most of them could have been fixed in 20 minutes with the right feedback.