Personalisation at Scale: 20 Hooks in 30 Minutes

The middle path between manual personalisation that caps at 10 people and fully automated outreach that reads like a machine wrote it

The goal is not personalisation or scale. It is both.

Manual personalisation produces excellent emails but caps out at 5 to 10 people per week. Fully automated AI outreach can reach 500 people but reads like what it is. Neither extreme is right for a solo operator building a services business.

The middle path is to use Claude to produce the raw material, apply human judgment to every output, and send only what you would be comfortable attaching your name to.

The One Real Line Rule

Every cold email you send must contain exactly one sentence that could not have been written about any other person on your list.

That sentence is the hook. Everything else in the email can follow a consistent structure. The hook must be unique.

When you have 20 prospects that means 20 different hooks. The system below produces them without spending 4 hours per week researching.

Step 1: Research Pass (10 minutes)

For your list of 20 prospects spend 30 seconds per person. Open their LinkedIn profile. Check their last 3 to 5 posts. Open their company website. Note one thing that connects to the problem you solve.

Record it in the Signal column of your CSV. It does not need to be polished, just the raw observation. Something like "Posted about their team doubling. Website still looks like it was built in 2019. Booking form is a PDF."

Do this for all 20 before moving to writing. Batching the research is faster than switching between research and writing.

Step 2: Batch Hook Generation (10 minutes)

Once you have 20 raw signals, run this prompt:

• I have 20 prospects I am writing personalised cold email hooks for.
• Using @EMAIL.md — write one opening hook of 2 sentences maximum for each
• person below.

Each hook must:

• Reference the specific signal I found and not a generic observation
• Sound like a person wrote it and not a template
• Not mention what I do or sell — stay in their world
• Be different from every other hook in this list

What I solve: [ONE SENTENCE FROM SALES.MD OFFER DESCRIPTION]

Prospects and signals:

1. [Name] — [Role] — [Company] — Signal: [YOUR RAW NOTE]
2. [Name] — [Role] — [Company] — Signal: [YOUR RAW NOTE]

...20 rows

Output format for each:

[Number]. [Name] @ [Company]: "[Hook text here]"

Flag any where the signal is too vague to write something specific —

I will research those manually.

Step 3: Human Review Pass (10 minutes)

Read every hook Claude generated. For each one ask four questions.

Would I say this to someone's face? If it feels strange to imagine saying it aloud, rewrite it.

Is it specific enough? "I noticed your website" is not specific. "Your booking form on the services page requires a PDF download" is specific.

Does it sound like me? Claude tends to write more formally than most solo operators speak. Simplify any sentence that feels polished in a bad way.

Is it accurate? Claude occasionally hallucinates details especially for LinkedIn signals. Verify that the thing it says you noticed is actually there.

Rewrite or remove any hook that fails these checks. A list of 18 good hooks is better than 20 mixed ones.

Step 4: Complete the Emails (5 minutes)

The bridge and the ask are largely the same across a campaign that targets the same problem for the same type of prospect. Write one standard bridge and one standard ask. Apply them to all 20 emails with the unique hooks already in place.

Your 20 first emails now look like this: unique hook that is different for all 20, plus consistent bridge, plus consistent ask.

This is the output that gets signal-based reply rates of 5 to 18% while being producible in 30 minutes per week.

Signals vs. Merge Fields: The Critical Distinction

Merge fields are not personalisation. They are just automation.

"Hi [NAME], I work with companies like [COMPANY] in the [INDUSTRY] space..." is a merge field. The prospect can see it was generated from a spreadsheet. It does not prove you looked at their business.

Signals are different. A signal is something you could only reference if you actually looked at this specific company.

The test: could this sentence have been in any other email you sent this week? If yes it is a merge field and not a personalisation. Rewrite it.

Batch Personalisation for Specific Campaign Types

For website audit campaigns visit 20 websites and run this analysis prompt:

I am about to visit 20 websites for a cold email campaign.

What I build: [YOUR OFFER FROM SALES.MD]
Target problem: [THE SPECIFIC ISSUE YOU SOLVE]

For each website below I will paste the homepage content. Identify one

specific observation that connects to my target problem — something a

non-technical business owner would recognise as accurate and feel about.

Do not mention loading speed unless it is extremely obvious. Focus on:

• missing functionality, friction in the flow, mismatch between what the
• site promises and what it delivers, features competitors have that this
• site is missing.

Website content: [PASTE HOMEPAGE TEXT AND VISIBLE ELEMENTS]

For LinkedIn post campaigns paste their recent post text:

A prospect posted this on LinkedIn [PASTE POST TEXT]. I want to reference

it in a cold email hook. Write 2 hooks that:

1. Quote or closely reference what they said
2. Connect to a gap or opportunity their post implies
3. Do not sound like I am complimenting them or have been stalking their feed

What Scale Actually Means for Solo Operators

You do not need to send 500 emails a week. You need to send the right 30 to 50 emails a week consistently over 3 to 4 months.

30 first emails per week with a 4-email sequence means roughly 120 emails in circulation at any time.

At a 7% reply rate that is 8 to 9 replies per month. At a 3% positive reply rate meaning people genuinely interested, that is 3 to 4 warm leads from cold outreach every month in addition to whatever inbound and referrals you already generate.

That is a meaningful addition to a services pipeline. Not hundreds of emails. Thirty a week done consistently.

Your Action Right Now

Run the batch hook generation prompt for your 20-prospect list. Review every output. Rewrite or remove anything that fails the specificity test.

You now have 20 personalised hooks ready to build emails around.