Lesson 3 of 10 · 8 min
Find the posts: by hand, or with Claude Code
The posts you want are the ones your buyer is reading and reacting to, about the problem you solve, published recently enough that a comment still gets seen. LinkedIn gives you three ways to find them by hand, about twenty minutes a day, and Claude Code gives you a fourth that does a month of finding in one sitting. The video uses the fourth. Both are below.
1. Search posts, filtered to this week
Type a phrase your buyer would use into the LinkedIn search bar, press enter, then choose Posts and set Date posted to past week. Not your jargon, theirs: a Shopify owner does not search “retention strategy,” they write “abandoned cart” and “repeat customers.” Try five or six phrases. Save the ones that return real conversations rather than link dumps. The results are sorted by engagement by default, which is what you want: a post with forty comments has forty people reading the comments.
2. Follow the thirty accounts your buyers follow
Every niche has twenty or thirty people whose posts your buyers read. Find them by looking at who your existing clients follow, and who is commenting thoughtfully on the posts from step one. Follow them, and use the bell on their profile to be notified when they post. A comment in the first hour of a post from one of these accounts is worth ten comments on dead posts, because the author’s audience is arriving at the same time you are.
3. Read the comment sections
The comments under a busy post are a list of people with the problem, in their own words, today. Some of them are better targets than the post itself. Reply to those too; a reply to a comment notifies its author directly.
The daily routine
- Open notifications. Comment on any post from your thirty that is under two hours old.
- Run your saved searches, past week. Pick five posts with real discussion under them.
- Read each post properly. Read the comments. Then write.
- Log every comment in the sheet from lesson nine before you move on.
4. Have Claude Code pull them into a sheet
This is what the video does, and it is where the five hundred comes from. Open Claude Code and give it one prompt with three parts: the topic in your buyer’s words, the number of posts, and your one-page offer document from lesson two. His, near enough: “Scrape me 500 LinkedIn posts where people are talking about GTM and cold outbound, and for each one write a comment reply: two short paragraphs of value on what the post actually says, then a free call to action tied to it.” Ask for a sheet with one row per post: author, link, the post text, the draft comment. You get a month of targets in an afternoon, already read for you, each with a first draft that a person then fixes and posts (lesson six).
Be careful with it. LinkedIn’s terms say no scraping and its filter is good at spotting a logged-in account that fetches hundreds of pages at machine speed; accounts get restricted for it, sometimes without warning. So: do not run it from the account you comment from, keep the number in the hundreds rather than the thousands, and treat the sheet as a reading list, not a posting queue.
Checked The scrape is the source’s, kept whole; the caution is ours.
In the video the posts come from a prompt to Claude Code: “scrape me 500 LinkedIn posts where people are talking about GTM and cold outbound,” and the result is a sheet with the post and a drafted comment side by side, which he shows on screen. LinkedIn’s User Agreement, section 8.2, item 2, says members will not “use software, devices, scripts, robots or any other means or processes ... to scrape or copy the Services.” The video does not mention that. It is the reason for the paragraph above, and the reason the three by-hand routes are on this page too.
Spotted something wrong in this lesson? Tell us, and we will check it and fix it.