A social media content calendar prompt produces usable plans only when it starts with themes, not with dates. Give a model four named content pillars and a fixed post count per pillar, and it will hand back a quarter of posts with a mix you can defend. Ask it for "a month of content" and it will hand back thirty variations of the same idea.
The reason is that a calendar is a distribution problem, while a pillar is a topic problem. Models are good at distributing a set they are given and poor at inventing a set that has structure. This guide gives you the three prompts that run in sequence, a worked quarter for a solo creator, and a review checklist to run before anything is scheduled.
Name the four pillars before you generate a single post
Four pillars is a working default, not a law. Three is enough for a narrow niche, five works for a brand with two audiences. The number that matters is the one you can sustain without repeating yourself, and the test is whether each pillar answers a different question your audience asks.
You are a content strategist for [channel or business]. My audience is [who they are and what they already know]. What I do or sell is [offer, product, or subject]. My realistic production capacity is [posts per week and hours per week]. The platforms I use are [list platforms]. Generate four content pillars. For each pillar give me: a short name, the one question this pillar answers for my audience, three topic seeds specific enough to write from, the post formats that suit it, and one thing this pillar must never cover so it stays distinct from the others. Keep the pillars visibly different from each other. Do not invent facts about my business, my results, or my clients. If you need information I have not given you, write [CONFIRM: detail needed] in place of the guess and list those items at the end.
Fill in these details
- Audience: Describe them by what they already know and what they are trying to do, not by a demographic label.
- Offer: What a new follower would actually buy or watch. Vague inputs produce vague pillars.
- Production capacity: The number here constrains everything downstream. Overstating it produces a calendar you abandon in week three.
- Platforms: Format choices differ sharply between a short video platform, a text feed, and a professional network.
- Non-overlap note: The exclusion line per pillar is what stops the model from quietly folding two pillars into one.
If the four pillars come back overlapping, do not generate posts yet. Rewrite the pillar names by hand and re-run. Everything downstream inherits this structure.
Generate ninety post ideas that do not repeat themselves
Ninety is a planning ceiling for a quarter, not a production quota. Treat the output as a stock of options you will cut roughly in half.
Using the four pillars below, write [90] post ideas for one quarter. Pillar names and rules: [paste the four pillars]. For each post give me: pillar, the specific angle or question it addresses, the format, a working title under ten words, and the one idea it delivers that a competitor could not. Spread the posts across the four pillars as evenly as the rules allow, but do not force a weak post to balance a total. Vary the angle across the quarter. If two ideas would open with the same sentence, rewrite one. Do not invent statistics, testimonials, prices, results, or events. Do not use a trend, holiday, or platform feature I have not listed. Mark any post that depends on a fact I must supply with [NEEDS FACT]. Return a markdown table with columns: number, pillar, angle, format, working title, unique value, facts needed.
Fill in these details
- Post count: Ninety works when you plan three months across several pillars. Lower it if one platform is your only channel.
- Angle: State it if you care. "Contrarian take", "worked example", "myth correction", and "behind the decision" produce visibly different posts from the same pillar.
- Formats: Only the ones you can actually shoot, record, or write this quarter.
- Facts needed column: Keeping the gaps visible is what stops invented specifics from reaching a published post.
Turn the idea list into a calendar you can fill
The idea list is unordered. This prompt does the scheduling work: spreading pillars across weeks, clustering formats so production batches together, and keeping one evergreen slot open.
Here is my quarter of post ideas: [paste the table]. Build a scheduling plan that respects these constraints. Posting days: [days]. Maximum posts per week: [number]. No more than [number] posts on the same day. Keep at least one evergreen post per week and hold [number] slots as flexible, to be filled with what performed. Group posts of the same format on nearby days so I can batch the work. Output a markdown table with columns: week, day, pillar, format, working title, channel, repurposes into, and status set to "draft". Then output a short second table showing how many slots each pillar and each format received. If the constraints cannot all be met, say which one you relaxed and why. Do not invent posting dates, platform features, or holidays. Do not fill the flexible slots.
Worked example: one quarter for a solo creator
A solo creator teaches small business bookkeeping, posts four times a week, and can shoot short video but cannot appear on camera every week. Two of the four pillars are video-led and two are text-led, so the batching rule matters.
| Week | Day | Pillar | Format | Working title | Channel |
|---|---|---|---|---|---|
| 1 | Tue | Reconcile without panic | Short video | The three numbers to check first | Short video platform |
| 1 | Thu | Reconcile without panic | Carousel | Month-end checklist, printable | Text feed |
| 1 | Sat | Client stories | Text post | What I stopped charging for | Professional network |
| 1 | Sun | Tooling reality | Short video | The spreadsheet that broke | Short video platform |
| 2 | Tue | Tooling reality | Carousel | Six fields your ledger must have | Text feed |
| 2 | Thu | Client stories | Text post | A fee I raised and what happened | Professional network |
| 2 | Sat | Reconcile without panic | Short video | Bank match errors, by cause | Short video platform |
Two things in that table are worth noticing. The video posts cluster on the weekend, so one filming session covers several. And week two holds an open slot, which is where the best-performing week one post gets repeated in a new form rather than replaced with something weaker.
Review the calendar before you schedule anything
1. Check the pillar mix by count, not by feel
Read the second table the scheduling prompt returned. If any pillar has more than roughly a third of the quarter, something has drifted, usually because one pillar had richer topic seeds than the others. Add seeds to the thin pillar and re-run rather than editing dates by hand.
2. Check the production load against a real week
Take one normal week, not the best week. Count the distinct production tasks: filming, editing, writing, and replying to comments. If the calendar needs more hours than you actually spend on this, cut the lowest-performing format rather than the lowest-performing topic, because the topic may just need a different format.
3. Trace every number back to you
Open each row and find the source of any figure, price, date, or claim. If a figure came from the model, it is unverified and must be replaced with your own or removed. A calendar makes errors easy to miss because it looks finished.
Common content calendar mistakes
- Starting from dates instead of pillars. The model fills gaps with the nearest familiar format, which produces a calendar of variations.
- Treating ninety posts as a quota. A planning ceiling you cut from is more useful than a target you miss.
- Counting repurposed posts as new posts. One shoot can supply four posts. Label derived posts so you can see the real output.
- Using one pillar to fill a slow week. If a pillar has no fresh angle, leave the slot flexible instead of repeating.
- Letting the model assign dates or platform features. Promotions, formats, and seasonal moments change. Supply them yourself from current guidance.
Frequently asked questions
How many posts should a content calendar plan for?
Plan more than you will publish and cut later. A quarter of roughly twice your realistic weekly load gives you options, because a fixed calendar with no slack turns one bad week into a broken streak.
Can the same pillar prompt work across several platforms?
Yes, if you pass the platform list and constrain formats in the idea prompt. The pillars stay constant, but a pillar that reads well as a carousel often needs rewriting to work as a short video or a text post.
Should I let AI choose my posting dates?
No. Give it your posting days and post limits, then check any seasonal or promotional dates against current platform guidance yourself, because those change faster than the prompt.
Plan one quarter, then reuse the structure
Take one pillar prompt, run it against your own offer, and only then generate ideas. Build the calendar, cut it to what you can produce, and keep the flexible slots. Next quarter you reuse the same four pillars and replace only the topic seeds, which makes planning a fraction of the work. For more social media prompts built around a repeatable planning system, browse the library.