Patch that changed everything (7/11)
So, if previous logic wasn't enough, we can prove it differently, using another hypothesis: Since the developer changed the text, and we know the approximate time of the patch, we can compare reviews made before the change with ones made after. This, by our hypothesis, should show the relative decrease in spike height (or volume), potentially removing it completely, if the text was the sole reason. But since only the text was changed, and the link to the page wasn't, we expect some of the spikes to stay. It is visually clear that the hypothesis survived again: Before the change, we had 2250/8350 = 27% of "added" reviews from spikes! After the change, within the similar time window (~2 days), we had 550/5400 ~ 10% of "added" reviews from spikes. After that, for 2 weeks+ it stayed at the same rate of 10%, sloooowly decreasing. This, honestly, was a surprising result for me. A 17% difference just from changing the text? I guess this is why a Call-to-action is so important, but more on that later, in the conclusion section. So far we've proved that having an NPC with that text and mechanic substantially affected the number of reviews, but we have more!
0/0