Claude read 12 of my newsletters and told me what wasn't working
Claude read 12 of my newsletters and told me what wasn't working.
My open rates were fine. My click rates were fine. But something wasn’t quite there, and I couldn’t name it. So I stopped guessing and handed the problem to Claude.
I gave it 12 back issues, my average subject line performance, and a rough sense of who my readers are. Then I asked it to find what I couldn’t see.
What came back was uncomfortable. My article titles were doing a decent job. My subject lines were not. My SEO titles were barely trying. And my social preview copy was written for people who already knew me, which is the opposite of how discovery works.
This article is the prompts I used, in the order I ran them, with sample inputs. You can copy every one of these straight into Claude and run them against your own newsletter.
Start with a baseline: what are your titles actually doing?
Before you fix anything, you need to know what’s broken. This prompt pulls the patterns out of your existing titles so you can see them clearly.
Prompt 1: The title pattern audit
What it does: Surfaces the structural habits in your existing article titles so you can spot what’s repetitive or weak.
When to use it: Before you write a single new title. This is your starting point.
The prompt:
Here are [NUMBER] article titles from my newsletter, [NEWSLETTER NAME]:
[PASTE YOUR TITLES HERE]
Analyse these titles and tell me:
1. What structural patterns repeat (e.g. “how to X”, “X things that Y”, question formats)
2. Which titles lead with benefit vs. curiosity vs. fear vs. identity
3. Which titles would only appeal to people who already know what I write about
4. Which titles could appear on any newsletter in my niche without sounding different
5. Give me a short summary of what my titles collectively signal about the newsletter
Be specific. Use examples from my list.
How to use it:
Paste in 10 to 20 recent article titles (more is better)
Read the patterns section carefully. That’s where the useful stuff is
Note which titles Claude flags as “could be anyone’s.” That’s your problem list
Example input:
Here are 14 article titles from my newsletter, [TITLE]: How to write emails that convert, 5 ways to grow your list faster, Why your open rates are dropping, The one email metric you’re ignoring...
What you’ll get: A breakdown showing you which titles are structurally identical to each other, which lead with the reader’s problem vs. your expertise, and a plain-English summary of what your titles collectively promise.
Advanced note: Ask Claude to rank your titles by “specificity” after the audit. Vague titles cluster at the bottom. That ranking usually shows a clear line between your better-performing content and the stuff that never got traction.
Now do the same for subject lines
Article titles and subject lines are different things doing different jobs. Titles need to hold up as evergreen content labels. Subject lines need to get opened by someone scanning an inbox at 7am. Most newsletters write them the same way. That’s the mistake.
Prompt 2: The subject line weakness finder
What it does: Identifies where your subject lines are losing opens before anyone reads a word.
When to use it: When you suspect your subject lines aren’t matching the quality of your content.
The prompt:
Here are [NUMBER] email subject lines from my newsletter. Next to each one, I’ve included the open rate.
[PASTE YOUR SUBJECT LINES AND OPEN RATES]
My average open rate is [X]%.
For each subject line, tell me:
1. What type of subject line it is (curiosity gap, direct benefit, personal/conversational, news hook, etc.)
2. Whether the subject line would make someone feel like they’d miss something if they didn’t open it
3. One specific reason it might underperform
4. A rewritten version that keeps the same topic but increases urgency or specificity
Then give me an overall pattern: what type of subject lines consistently perform above and below my average?
How to use it:
Pull your last 20 to 30 subject lines from your email platform
Include open rates if you have them. If you don’t, still run the prompt and skip the pattern analysis at the end
The rewrites are worth reading even if you don’t use them directly. They show you the gap
Example input:
“This week’s best AI tools” - 31% “I almost missed this” - 44% “How I saved 3 hours on research” - 52% “The prompt I use every Monday” - 48%
What you’ll get: A line-by-line breakdown with a rewrite for each subject line, plus a pattern summary telling you which subject line types work for your specific audience.
You’ve seen what the audit looks like on the surface. The next six prompts go deeper: fixing your SEO titles, rebuilding your social previews, and running a full reader-language audit that shows you the exact words your audience uses to describe their own problems. That’s where most newsletters lose the plot.
[Upgrade to access Prompts 3 through 8, plus the bonus template: a complete newsletter title system with one reusable prompt for generating 10 title variants from any article.]