How to build a social media content workflow that survives a busy week.

The blank page is a sourcing failure, not a writing failure. If you sit down on Monday with nothing captured, you will produce generic advice, because generic advice is what is available when you have no specifics to hand.

Keep one running list and add to it the moment something happens: a feature shipped, a support question asked twice, a decision reversed, a number that surprised you, a competitor comparison a prospect raised. Ten seconds each, no formatting. The list is not content, it is raw material, and the only requirement is that everything on it actually happened.

A week of specifics beats a month of frameworks. Nobody needed another post about consistency; they might genuinely want to know why you removed a feature.

Planning and writing are different kinds of thinking, and doing them together is why both take so long. Planning is selection: of the fifteen things on the list, which four are worth saying, in what order, and to whom.

Do it before the week starts, when you can still see it whole. A plan made on Sunday commits you to four decisions; a plan made each morning asks you to make one decision daily under time pressure, which is the same job done worse five times.

Give each planned item a one-line angle rather than a topic. \u201cPricing\u201d is a topic. \u201cWhy we moved X off the free plan and what it cost us\u201d is an angle, and it is almost written already.

This is where most workflows quietly become cross-posting. One caption is written, pasted everywhere, and performs on none of them, because the platforms want genuinely different things. A thread that works on X is a wall of text on LinkedIn; a post that lands on LinkedIn reads as corporate on Bluesky; a link that is fine on Threads is dead weight where the algorithm penalises it.

The fix is not writing four unrelated posts. It is keeping one source idea and changing what the platform actually cares about: the hook, the length, the formatting, and what you want the reader to do. The claim stays identical. The delivery does not.

Practically: write the version for your strongest channel first, at full length, then adapt down. Adapting a good long draft is fast. Expanding a thin short one is not.

Writing and judging are different jobs and should not happen in the same hour. The draft that felt sharp while you were writing it is the one you are least able to assess, because you are still inside the sentence.

Review the whole week at once, deliberately later. Two questions, in this order: is it still true, and is it worth someone's attention. Truth first, because fluent copy hides stale facts well. An outdated price, or a claim about a feature that shipped differently, reads perfectly and is still wrong.

Keep a note of why you reject things. Repeated rejections for the same reason are the most useful output of the whole process; they tell you what to fix upstream rather than in this week's post.

Move the approved week onto a calendar in one action, then treat delivery as something that will occasionally break. Tokens expire, platforms have outages, media gets rejected, rate limits arrive at inconvenient minutes.

The thing that turns a failed post into a missed month is a failure you do not notice. Whatever you schedule with, make sure a failure is visible and the draft survives it, so that recovering means retrying an approved post, not rewriting one from scratch and re-deciding whether it was any good.

The five steps above work with a notes app, a document, and any scheduler. Plenty of people run them exactly that way, and if that is working, it is working. No tool improves a process you are already following.

Cadencz exists for the part that is genuinely tedious to do by hand: keeping the product context so step three does not start from a blank prompt each time. It reads your site once and stores what you sell, to whom, and how you write, then uses that to turn a step-one item into separate native drafts for each destination.

In workflow terms: it holds the source list, plans the week from it, drafts natively for 7 of its 9 live channels (X, LinkedIn, Bluesky, Threads, Mastodon, Telegram, and Discord), holds everything in one review pass, and schedules what you approve. Nothing publishes without that approval. The free plan runs the whole loop once so you can judge the output before paying for it.