Over the weekend, Li Banjia's "mini imperial court" worked almost without a break. (Saturday, Sunday, Monday, and Tuesday daily stipend +2 taels, the [Imperial Treasury] balance reached 807 + 2 * 4 = 815 taels. Imperial Majesty 32.)
After several days of accumulating initial data from the A/B test, a massive wave of information surged into the makeshift backend built by Zhang Chen. Starting Monday, the three of them began intensive data analysis and weekly report drafting in a temporarily borrowed small meeting room at the company.
True to his title as "Ministry of Works clerk," Zhang Chen wrote a script that clearly separated the two sets of test data and performed some preliminary aggregated processing. Pointing at the raw data stream on the screen, he explained to Li Banjia and Wang Mengmeng, "Jia-ge, Mengmeng, judging from the backend, Strategy A (welfare-incentivized) indeed has a slightly higher initial click-through rate and page arrival rate than Strategy B (content-engagement), by about 15%. However, the average page stay duration for users under Strategy A is nearly 30 seconds shorter than Strategy B, and its bounce rate is also somewhat higher."
Chapter 17