MAGEEEK

MAGEEEK helps Japanese students apply to universities abroad. Having completed that process myself, I joined the team to mentor current applicants and analyze marketing performance across X, YouTube, and the company website. Over four months, the share of X posts exceeding 10,000 impressions increased from 7.6% to 18%.

I classified twelve weeks of posts by topic, format, and the way each post opened. To avoid adjusting the categories to fit the results, I defined the criteria before beginning the analysis. I compared responses using a combined measure of likes, shares, and replies, and reviewed saves as a sign that readers wanted to return to the information later. This approach separated posts that simply reached large audiences from those that prompted meaningful action. I also found that the top ten posts accounted for more than half of all views, showing that results were concentrated in a small number of posts.

The analysis showed that performance depended less on emoji use or posting time and more on how information was presented. High-performing posts repeatedly used one of four approaches: a specific number, a view that challenged conventional thinking, a memorable statement, or a clearly organized list. Rather than stopping at a report, I turned these findings into eight post concepts for a 30-day trial, giving the team a plan it could implement and evaluate immediately.