How to personalize cold outreach so it does not read like a template
Everyone now recognises the email that says "I noticed your company" and then says nothing about the company. Real personalization is not a sentence you add. It is the reason the message exists.
Revio AI3 min read
The short version
- Personalization is not inserting details. It is writing a message that would make no sense sent to anyone else.
- The test: swap in a different company name. If the email still works, it is not personalized.
- Short beats clever. One observation, one question, no attachment.
- Ask a question they can answer in one line, not for thirty minutes of their time.
Why merge fields stopped working
For years, cold email personalization meant variables: first name, company name, maybe city. That worked while it was unusual. Now almost every sales email has them, and readers pattern-match in the first line. "Hi Sam, I came across Acme and was impressed" says, in effect, that this was sent to hundreds of people.
AI writing tools made it worse before they made it better. A model asked to "personalize" with nothing to go on produces flattering filler about the company's "impressive growth" and "commitment to quality". It reads as personal and says nothing.
The swap test
Three levels of personalization
| Level | Example | Effect |
|---|---|---|
| Cosmetic | Hi Dana, I love what Nova Health is doing. | Read as a template |
| Researched | I saw Nova Health has seven clinics in the metro area. | True, but so what? |
| Relevant | You have had two front-desk roles open since July across seven clinics. Who is picking up the calls meanwhile? | A reason to reply |
The jump from researched to relevant is the one that matters. A researched fact shows you looked. A relevant fact connects what you found to a problem they probably have, which is the only reason a busy person answers a stranger.
Where the relevant fact comes from
It comes from research, and specifically from signals that a business may have the problem you solve. How to find high-intent prospects lists the signals worth using. The short version: things a business has published about itself recently, such as hiring, expansion, pricing changes, and recurring themes in its reviews.
A structure that works
Most good cold emails fit in four short parts:
- The observation. One specific, true thing about their business. No compliments.
- The implication. What that probably means for them, stated plainly.
- The relevance. One line on what you do, only as it relates to that implication.
- The question. Something they can answer in a sentence.
Saw you published retainer pricing last month, which usually means volume matters more than fit from here. Right now it looks like the two partners are doing all the selling between client work. We run the first conversation for agencies in that spot. Who is handling new enquiries at the moment?
Subject lines
Keep them short, plain and specific. The goal is to look like a normal email from a person, not a campaign.
| Avoid | Better |
|---|---|
| Quick question!! | The two front-desk roles |
| Grow your revenue by 300% | Quote callbacks in Bixby |
| {FirstName}, an opportunity for {Company} | Retainer pricing |
Misleading subject lines are not just ineffective. In the US, deceptive subject lines are prohibited under CAN-SPAM. More on the rules in cold email deliverability rules.
What to leave out
- Compliments you cannot back up. "Impressed by your work" from a stranger is noise.
- Your company history. Nobody needs your founding story in a first email.
- Attachments and images. They hurt deliverability and nobody opens a PDF from a stranger.
- A calendar link as the only option. Asking for thirty minutes before there is any interest is a big request. Ask a question first.
- Fake familiarity. "Following up on my last email" when there was no last email.
Personalizing follow-ups too
Follow-ups are where personalization usually collapses into "just bumping this". Each follow-up should add something: a different angle on the same problem, a relevant example, or a clean way to say no. How automated sales follow-up works covers how to do that at volume.
Doing this at scale
The honest difficulty is that relevant personalization requires research, and research takes time. That is exactly what software is now good at: reading every business properly and writing from what it found. The rule to hold any tool to is simple. If research found nothing specific, it should not send anything. A generic message sent under your name costs more than it earns.
Revio writes each email from its research on that business, and holds off on businesses where it has nothing specific to say.
Common questions
How long should a cold email be?
Short enough to read on a phone without scrolling. Four to six sentences is plenty for a first message.
Is using AI to write cold emails a problem?
Not in itself. The problem is AI writing with nothing to go on, which produces generic flattery. AI writing from real research about a specific business can be more relevant than a rushed human draft.
Should I mention a competitor I noticed they use?
Usually not in the first message. It can come across as presumptuous. Stick to what they have published about their own business.