You will be able to compare posts in a way that finds real patterns instead of chasing one lucky result.
Mei Ling posts a video of a cake shaped like a cartoon character, and it reaches far more people than anything she has posted before. Excited, she spends the next month making nothing but character cakes. Her reach drops back to normal, then lower, and the enquiries she was getting from her party-planning videos dry up.
She had treated one result as a rule. It happens to almost everyone who starts reading their analytics, and this lesson is about how to avoid it.
The reach of any single post depends on many things you do not control: who happened to see it first, what else was in the feed that day, whether one person with a large following shared it. A post that goes far once may have been a good post, or lucky, or both. You cannot tell from one result.
What you can trust is a pattern. If several of your strongest posts share something, such as a pillar, a format, a type of hook or a length, that is worth acting on. If only one post has it, wait and test it again.
So when a post does unusually well or unusually badly, note it, ask what might explain it, and hold off on changing your plan until you see the same thing again.
Comparing every post with every other post mixes too many differences together. A talking-to-camera opinion video and a hands-and-process cake video differ in pillar, format, hook, length and subject. If one does better, you have no idea which difference mattered.
Compare posts that share most things, such as two videos in the same pillar, two in the same format, or two with the same hook pattern from lesson 3.2. The fewer differences between them, the more you learn from comparing them.
It also helps to compare rates rather than raw totals. A video with more views will usually have more saves just because more people saw it. Divide by views to compare fairly. As an example, Priya's first video had 4,000 views and 40 saves, which is 10 saves per 1,000 views. Her second had 1,200 views and 30 saves, which is 25 per 1,000. The second video reached fewer people, but the people it reached found it much more worth keeping, and that is the kind of finding a raw total hides. All four figures are examples.
Do the same with watch time, by comparing it with the video's length, as lesson 9.1 showed.
If you want to learn what works, change one thing between posts and keep the rest steady. Same pillar, same format, same length, but a different type of hook. Or the same hook pattern but a shorter video. Then compare.
If you change the hook, the format, the posting time and the topic all at once, and the next video does better, you cannot say why. You learn nothing you can repeat.
This is slower than changing everything at once, and it can feel less exciting. Over a few months it teaches you more about your own audience than any article about what works on social media. Keep a note of each test in your sheet: what you changed, what you kept the same, and what happened.
Farah ran this on her hooks. For four weeks she kept the same format and length and alternated between mistake hooks and question hooks. The mistake hooks held viewers longer, on average, in three weeks out of four. That was a pattern worth acting on, and it came from four weeks of patience.
To compare like with like, you need to know what each post was. Set up a sheet with one row per post and these columns: date, pillar, format, hook type, length, reach, average watch time, saves, shares, profile actions, and messages or enquiries from your log. Fill in the tags when you post, and the numbers a few days later.
After ten posts, sort and filter. Which pillar gets the most saves per view? Which format holds attention longest? Which hook type leads to the most profile actions? You are looking for one pattern you can act on, and a full theory of your audience can wait.
Social analytics only see what happens inside the app. Working out how social activity leads to sales across your website, email and other channels is called attribution, and it is covered in Marketing analytics: GA4, attribution and testing.
In the activity you will put your last ten posts in a sheet, tag each by pillar, format and hook type, and look for one pattern.
Put your last ten posts in a sheet, tag each by pillar, format and hook type, and look for one pattern.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).