LinkedIn post ideas from email marketing data that sound like you
The best founder and operator posts start from something real you saw in your own list. Here is how to find it, shape it and post it without giving away client or customer data.
The best LinkedIn post ideas from email marketing data come from one thing you saw in your own list: a test you expected to lose, a flow you changed, or what new subscribers wrote back. Write that observation down with its number, say what you did and what happened, and end with one call to action that fits. That is a full post, and it sounds like you because it happened to you.
This guide shows you how to find those observations, how to shape them into a post and a short note that goes deeper, and how to keep client names and customer data out of it. You can do all of it by hand, or run it through the free / skill in Claude. It is workflow 6 in Claude for email marketing: the 6 workflows.
One real observation from your email program (with the number if you have it), one line on who reads you on LinkedIn, and two posts you already wrote that sound like you. Nothing else. If you do not have the observation yet, the first section tells you where to look.
01LinkedIn post ideas from email marketing data: where to look
Most LinkedIn content for ecommerce founders reads the same because it starts from opinions. Your email program gives you something most people posting do not have: a record of what real subscribers did. Every send and every reply is a small experiment you already paid for.
Open your email platform and look in these places first:
- A/B tests. Especially the ones where the version you expected to lose won.
- Flow changes. Anything you edited in a welcome, abandoned checkout or post-purchase flow, with the numbers from before and after.
- Replies. What new subscribers write back when an email asks them a question.
- Unsubscribes and complaints. Which email caused a spike, and what you think it was about.
- Signup forms. A change to the offer, the copy or the timing of a popup, and what happened to signups.
Keep a running note. Each time something in the account surprises you, write one line: the date, what you saw, and the number. That note becomes your backlog of post ideas.
02How to spot a postable observation
Not every number is a post. Run each line in your note through three tests. If it fails one, leave it in the note and wait for more data.
Specific. It names one email, one flow or one change. "Our welcome flow" is too broad. "Moving the discount from email 1 to email 3" is specific.
Surprising to you. You expected something else. If you saw it coming, your readers probably did too. Your surprise is the hook, so say what you expected.
A number or a before and after. A count, a rate, or a clear change over time. It can be round or relative, as long as it is true. Without one, the post is an opinion.
One more rule: one observation per post. If you have three findings, you have three posts. Stacking them makes each one weaker and the post longer.
03Four observation-to-angle examples
These four are illustrations, made up to show the method. They are not client results, benchmarks or predictions of what you will see. Swap in your own observation and your own number.
Observation: a plain-text welcome email beat the designed version on clicks. Angle: "I spent a week on the designed one and the plain one won." The post is about testing your own taste against your list. Keep the claim to this one test.
Observation: moving the welcome discount from email 1 to email 3 changed who bought first. Angle: what you give away first teaches people how to buy from you. Say what you changed, what before and after looked like, and that it is one store.
Observation: you asked new subscribers why they signed up, and roughly a third said they were buying a gift. Angle: your subscribers wrote your positioning for you. Show the question you asked and what you changed on the site or in the flow because of the answers.
Observation: you removed subscribers who had not opened in months, the list shrank, and revenue per send held steady. Angle: list size means little if you cannot reach the people on it. Give the relative change, and say what you still do not know.
Each angle claims less than the observation might support, on purpose. A post that claims less is easier to trust.
04How to structure the post
Use the same four parts every time. It keeps you honest, because each part has to be true on its own.
- Observation. What you saw and what you expected, in one or two lines. This is the hook, so the surprise and the number go here.
- What you did. The change or the test, in plain words. Enough detail that another operator could repeat it.
- What happened. The result with its number, and how long you watched it. Say so if the sample was small.
- What you would do next. Your next test, or what you would tell someone trying the same thing. This is where your judgment shows.
Then add one call to action that matches the post. An honest CTA asks for the thing the post naturally leads to:
- If the post is a finding, ask: "Have you seen the same in your list?"
- If the post leaves a question open, ask readers what they would test next.
- If you wrote a longer note on it, point to it once, in one line.
One CTA per post. When a post is about a lesson, a sales pitch at the end reads as the real reason you wrote it. Save the offer for posts that are about the offer.
The short note is the deeper version: the full setup, the numbers you left out of the post, what you got wrong, and the next test. Three to five short sections, in your newsletter, a LinkedIn article or on your site.
05Privacy: what to leave out
Your email data is about real people, and often about a client's business. Before anything goes public, go through this list:
- No client names without written permission. "A skincare brand I work with" is enough. Also remove details that identify them anyway: the exact product, the city, a launch people would recognize.
- No customer data. No names, email addresses, order details or screenshots of replies that show who sent them. If you quote a reply, paraphrase it or get permission, and strip anything that points to the person.
- Round or relative numbers when the real ones are sensitive. "Roughly a third" or "about double" is still true and keeps revenue private. Never round in the direction that flatters you.
- Crop screenshots. Dashboards show account names, list sizes and revenue in corners you forget about.
LinkedIn's Professional Community Policies say: "Do not reveal others' personal or sensitive information." The same page asks members to share "information that is real and authentic" and says "Don't do things to artificially increase engagement with your content." A post built on your own observation, with an honest CTA, fits all three.
06The prompt
You can write the post by hand with the four parts above. If you want help, the free / skill turns a real observation into a LinkedIn post, a note outline and one honest CTA, in your own voice. It arrives by email when you join the Margarita Creative newsletter. Without the skill, the same prompt still works in Claude: delete the first sentence and send the rest.
Use /list-first-content. Here is something real from my email program: [OBSERVATION, WITH THE NUMBER IF YOU HAVE ONE]. My audience on LinkedIn: [WHO]. Two posts that sound like me: [PASTE THEM].
1. Write one LinkedIn post built on this observation, in my voice.
2. Outline a short note that goes deeper.
3. Give one honest call to action that matches the post.
Use only the facts I gave you. No invented results and no hype.The two sample posts matter most. They are how Claude picks up your sentence length, your openings and the words you never use. Pick two you wrote yourself and still like. If you have none yet, paste two emails you wrote to your list.
The last line is the guardrail. If the draft adds a number, a percentage or a result you did not give it, that line was ignored. Delete the claim and ask again.
07How to check the output
- Every number is one you gave. Go through the draft line by line. Any figure you do not recognize comes out.
- No client names without permission. Also check for details that identify them anyway.
- Read it aloud. It should sound like you talking to another operator. Rewrite any sentence you would not say out loud.
- One observation, one CTA. If the draft drifted into a second finding or a second ask, cut it.
- Reply to every comment in the first hour. The people who comment early are the ones most likely to talk shop, so be there while they are.
Finding where your list leaks and testing subject lines both produce the kind of before and after numbers that make good posts. The full series is in the pillar guide.
08Quick answers
What if I do not have a number?
Use a before and after instead, such as what subscribers replied before and after you changed a question. If you have neither, keep the observation in your notes until you do.
Can I post results from a client's account?
Only with their permission, and even then remove anything that identifies them or their customers. Without permission, describe the business in general terms and use relative numbers.
Should I round my numbers?
Round or use relative numbers when the exact figure is sensitive, such as revenue. Keep the rounding honest and never round in the direction that makes the result look better.
How long should the post be?
Long enough for the four parts and nothing more. If one part needs more than a few lines, move that detail into the deeper note.
Do I need the /list‑first‑content skill?
No. The four-part structure and the prompt work in Claude without it. The skill is free and saves you from writing the instructions each time.