In a nutshell
A/B testing is a way to publish version A and version B, which differ in only one thing, and compare which did better. It is widely used to improve ads and websites, and the idea works for social media posts too.
How to compare social media posts
With ads, you can show A and B to different people at the same time. Ordinary posts can't do that, so it is important to keep the conditions of the comparison as similar as possible.
| What you compare | A | B |
|---|---|---|
| Image | Just the dish | A person in the shot |
| Opening line | Conclusion first | Starts with a question |
| Time of day | 12:00 noon | 8:00 p.m. |
| Length | Short text | A slightly longer explanation |
Change only one thing at a time. If you change the image, the opening line and the time of day all at once, you can't tell which one made the difference.
Avoiding misreadings
A single comparison may just be a chance difference. A trending topic that day or good weather can swing reactions a lot. Repeat the same comparison several times and see whether the trend holds. For the numbers you compare, rates such as engagement rate, which are less affected by differences in impressions, work well. The basics of reading the numbers are covered in the basics of social media analytics.
Write down what you compared and what happened, even briefly. A note like "the image with a person got better reactions three times in a row" means you won't repeat the same test even if the person in charge changes. How to roll these notes into a monthly review is covered in how to build a monthly report. Fold what you learn into the rules for your next posts, then test one new thing, and keep that cycle going.
In SNS TOWN
SNS TOWN has no dedicated A/B testing feature, but Analytics shows a table comparing reactions per post with and without images or videos. For comparing times of day, you can use best posting times, which summarizes reactions by day of week and hour.
可免費試用 7 天，無需綁定信用卡。
免費開始
