How to test what works on your profile without guessing
Generic advice works on average. Your profile is not the average. The only way to know what works with your audience is to test in an organised way.
The fundamental rule
One variable at a time. If you change format, timing and topic in the same week and reach goes up, you do not know which of the three worked — and you cannot repeat it.
What is worth testing
- Format: video, carousel, single photo.
- Opening: question, strong statement, result first.
- Length: 15, 30 or 60 seconds.
- Timing: morning, afternoon, night.
- Topic: which of your pillars earns the most saves.
- Call to action: closed question, keyword, save request.
How to set up a test
- Pick one variable.
- Produce at least four posts with each option.
- Keep everything else the same.
- Compare the average, not the best post.
One post proves nothing. Four posts start to show a pattern.
Which number to look at
It depends on the goal. To grow, look at reach among non-followers. For relationship, look at saves and comments. For sales, look at profile visits and link clicks.
Write it down
A simple spreadsheet with date, format, topic, timing and the main numbers. In two months you will have more information about your profile than any outside advice could give.
When to stop testing a variable
When a pattern repeats three times. Then it becomes a rule for your profile, and you move to the next variable.
Watch for seasonality
Holidays, month-end and special dates change audience behaviour. Tests run in those periods can mislead.
Summary
Test one variable at a time, with four posts and average comparison. It is slow, but it produces knowledge that applies only to your profile — and that is the only kind that matters.