TL;DR: Treat social posts for a course launch as the top of a funnel that ends at your email list or waitlist, not at the checkout page. Plan three phases (pre-launch, a short open window, a close), write the facts once, give each post a tracked link, and let an AI agent reshape and queue the drafts. You still write the stories and answer the replies.
You spent two months recording the course. Launch day arrives, you post “Enrollment is open!” on three platforms, and the replies are a few congratulations and no sales.
That’s the usual result of a launch announced once to people who never heard the course was coming. Most launch guides agree on the fix: a pre-launch period, a short launch window, and a follow-up. Ruzuku’s course promotion guide goes further and says that, on its platform, the large majority of enrollments come from email subscribers and live workshop attendees rather than from social algorithms. That is a vendor’s own figure, so take it as a direction, not a law. But it points at the right role for social posts: get people onto a list or into a live session, then sell there.
This guide covers the social half. It’s about what to post when, per platform, and how to hand the repetitive drafting to an agent.
The three phases
| Phase | When | Social job | Link goes to |
|---|---|---|---|
| Pre-launch | 4-8 weeks before | Show your thinking, collect a waitlist | Waitlist or free lead magnet |
| Launch window | 5-10 days | Announce, prove, answer objections | Sales page (with a UTM link per post) |
| Close | Last 24-48 hours | Deadline, final answers | Sales page |
Guides differ on length. A Payhip guide and others describe similar pre-launch, launch and follow-up stages with varying timelines, so use them as a range and adjust to how long your own audience takes to decide.
Pre-launch: teach, don’t tease
Several guides suggest roughly four helpful posts for every promotional one. In practice that means posting pieces of the course itself: one lesson as a thread, a mistake students make, a before-and-after from a beta student. Put the waitlist link in your bio and in about one post in five.
A useful pre-launch post answers a question your future student actually has. If you can’t think of five, the course topic may be vaguer than you think, and that’s worth finding out now.
Launch window: one story per post
Don’t repeat “enrollment is open” every day. Give each post a different job:
- Open: what it is, who it’s for, what they can do after, price, link.
- Proof: one student result, with permission, in their words.
- Objection: “I don’t have time” or “I’m a beginner”, answered with specifics, such as how long a lesson really takes.
- Behind the course: a screenshot of a module or a story about why you built it.
- Live session: invite people to a free workshop or Q&A where you teach one part and answer questions.
Close: honest deadlines
If you set a deadline, keep it. Fake urgency that resets is easy to spot and costs trust. Post once the day before, once on the morning of close, and once about two hours before. Then post a short thank-you and, if the course stays open, say so plainly.
Per-platform adjustments
- LinkedIn: Works for professional or career courses. Write in first person, tell the student-result story, put the link in the first comment or the post itself depending on what you find reaches more people. Test it, don’t trust a rule.
- X, Bluesky, Threads: Short posts, one image, the date or deadline up front. Threads of a free lesson work well in pre-launch. Mastodon is similar, but add two or three CamelCase hashtags since that’s how strangers find you.
- Dev.to, Hashnode, Medium, Substack: For technical or writing courses, a long free lesson with an invitation at the end often does more than an ad. See auto-publishing to Hashnode and Substack.
- Pinterest: Good for visual, evergreen topics (design, crafts, planning). Pins keep sending traffic after the launch ends. See Pinterest auto-pin scheduling.
- Reddit: Many subreddits ban course promotion or limit self-promotion. Share useful material, answer questions, and mention the course only where the rules allow. Do this one by hand.
Give every post its own link
Add UTM parameters, documented in Google’s campaign URL guidance, so you can see which platform and which post got people to the page:
?utm_source=bluesky&utm_medium=social&utm_campaign=course-oct&utm_content=objectionChange source per platform and content per post. After the launch, compare. The weekly review workflow covers how to read it. If the objection post drove more signups than the open post, lead with objections next time.
Where it goes wrong
- The price or deadline changes and queued posts still show the old one.
- A student result gets posted without their permission.
- Refund or access questions arrive in replies and sit unanswered during the busiest week.
- Pre-launch posts sound like ads, so people stop reading before the launch.
Keep the plan in one file, re-read it the day before opening, and if a fact changes, delete the queue and rebuild it.
Handing the drafting to an AI agent
A five-post launch window across four platforms is twenty drafts. The mechanical part is reshaping; the decisions are yours.
Write a launch.md with the course name, one-line promise, audience, price, open and close times in UTC, the base URLs for the waitlist and the sales page, and 3-5 approved student quotes. Ask the agent to produce per-platform drafts with UTM links, then approve them.
With PublishPort, you log in to your own accounts once in the desktop client. Your AI agent connects through the relay, calls list_capabilities to see which platforms are connected, reads each tool’s -h output, and posts from your own logged-in sessions, with dry-run available first. See pricing for plans. If an agent has never posted for you before, run a throwaway test post first. For the write-once pattern, see repurposing one post across platforms; for the live-session part of the launch, see promoting a webinar.
Limits
- The desktop client must be running and logged in when a post is due.
- Platforms change their composers and rate limits, so check results instead of assuming success.
- The agent can’t judge whether a testimonial is accurate or whether your deadline is real.
- Replies, especially refund and access questions, should come from you.
Checklist
launch.mdwith UTC times, prices, links and approved quotes- Waitlist link in bio before pre-launch starts
- One UTM link per platform and per post
- Student permission for every result you share
- Reddit written by hand against subreddit rules
- Queue re-read the day before opening
- Time blocked for replies during the window
FAQ
How early should I start promoting an online course?
Guides range from a few weeks to a few months. Start when you can post something useful weekly and have a waitlist link ready.
Which social platform is best for promoting a course?
The one where your future students already spend time. Pick one primary platform first and add others only if you can keep up. Check your own link clicks rather than a general ranking.
How often should I post during a course launch?
Enough to cover each job (open, proof, objection, behind the scenes, close) without repeating the same text. For most creators that’s one post a day on the main platform and fewer elsewhere.
Should I link to the sales page or a waitlist?
Waitlist or lead magnet before launch, sales page during the window. Many creators sell mainly through email, so the social goal is often getting people onto the list.
Can an AI agent run my course launch posts?
It can draft and queue or send posts from your connected accounts. You decide the offer, verify facts and handle replies.