AI Guides › Playbooks
The Thirty-Day Post Ledger: Deciding What to Post Next From Your Own Numbers
By Nigel Guy · 5 min read
Most people decide what to post next by looking at the last post that did well and making another one like it. It feels like data-driven work, and it isn't: one post is an anecdote, and the headline number on it (usually reach or likes) is the number most easily moved by luck, timing and a bit of paid boost. A month of your own page data, laid out properly, will tell you something different, and often something less flattering.
The rule: judge topics and formats, never single posts, and only against a question you wrote down before opening the numbers.
What you need for the audit
- Thirty days of posts on one page or account (sixty if you post fewer than eight times a month).
- The platform's own analytics export or a screen-by-screen copy of the numbers. Menu names and export options on Meta, YouTube, LinkedIn and TikTok change often, and I could not confirm current ones for this guide, so open your platform's help centre (linked below) and look for "export" or "download data" in its insights area.
- A spreadsheet. Free options exist (Google Sheets, LibreOffice); check any paid plan's £ price at checkout.
- Fifteen minutes to write your question first.
The mechanism: the Post Ledger
The ledger is one row per post, with the same columns every month. Fill in the first five from the platform and the last three yourself.
| Column |
What goes in it |
Why |
| Date and time |
When it went live |
Spots timing effects |
| Format |
Text, image, carousel, short video, link, etc. |
Formats behave differently |
| Topic tag |
One tag from a list of four to six you choose in advance |
Lets you compare like with like |
| Reach (or views) |
People or views the platform reports |
Size of audience, not quality |
| Action count |
Saves, shares, comments, clicks: whichever you actually want |
Quality of response |
| Promoted? |
Yes or no |
Paid reach contaminates organic comparison |
| Action rate |
Action count divided by reach |
Makes big and small posts comparable |
| Outcome |
What happened off-platform: enquiries, sign-ups, replies |
The only column that is not a vanity number |
How to run it in five steps
- Write the question. One sentence, such as "Do my how-to posts bring more enquiries than my opinion posts?" If you cannot write it, you are about to go fishing.
- Fix the tag list. Choose your four to six topic tags before looking at results. Tagging after seeing a post's performance bends the answer.
- Fill the ledger. Copy the numbers across. Mark promoted posts and exclude them from the comparison (keep them in the sheet).
- Group, then compare. Average the action rate by topic tag, then by format. Use the median if one post is wildly bigger than the rest. Look at the count too: a tag with two posts is a hint, not a finding.
- Write one decision. "Next month: three how-to posts, one opinion post, test one carousel." Then write what result would make you change your mind.
A worked example (hypothetical)
Imagine a small UK bakery page that posted twenty times in a month. The question was "Do behind-the-scenes posts or product photos lead to more orders?" Tags: behind-the-scenes, product, offer, community.
Suppose the ledger showed the best-reaching post was a product photo that a local group shared, with plenty of likes and no orders. Behind-the-scenes posts reached fewer people but produced most of the "how do I order?" messages. The tempting move is more product photos, because that post won on reach. The ledger says the opposite: reach was an accident of a share, while messages tracked a different tag. The decision becomes "more behind-the-scenes, with an order link in each", and the product photo is treated as a one-off. These numbers are invented to show the method; they are not a benchmark for any real business.
Where audits go wrong
- Reading reach as success. Reach says how far a post travelled, not whether anyone cared.
- Too few posts per tag. Three posts can't distinguish a pattern from luck.
- Mixing paid and organic. A boosted post will beat an unboosted one for reasons unrelated to the content.
- Changing definitions mid-month. Platforms rename and redefine metrics (views, reach and impressions are not interchangeable). Note the definition you used and keep it.
- Ignoring the outcome column. If you cannot connect any post to an off-platform result, say so; that is a finding too.
- Optimising for the algorithm's mood. One month of data describes this month's audience, not a law.
What good looks like, and what this won't do
A good audit ends with one written decision, one test and one stop doing. It will not tell you why people responded, predict next month, or prove cause: a topic can look strong because you posted it on better days. Treat each result as a prompt for the next experiment, and repeat the ledger monthly so the pattern has to survive more than one month.
What to skip
- Posting-time "best hours" charts from small pages; your own ledger is more reliable than a general chart.
- Follower count as a scorecard.
- Re-tagging posts after the fact to make a theory fit.
- Third-party dashboards you have to pay for until the free export has failed you.
Guardrails
- Use only data from your own page; don't scrape other people's accounts.
- If you are comparing many metrics, expect one to look good by chance. Stick to the question you wrote.
- Customer messages and enquiries may contain personal data: keep the ledger to counts, not names or message text.
- Platform metric names and export paths change; check the current help page before relying on a menu path from any guide, including this one.
- Nothing here is a statistical test. For decisions with real money behind them, run a longer test or ask someone who analyses data for a living.
Sources
All 751 AI guides · JulieMango plans from £17/mo