Video narration
Which distribution fits your problem: the Poisson distribution or the Normal distribution? Imagine comparing notifications with heights. How many notifications arrive in an hour? That is a count: zero, one, two, and so on. The Poisson distribution models event counts in a fixed interval, like an hour, when the average rate is stable and events are independent. People's heights are different: height is continuous, with many values clustered around an average. That is where the Normal distribution helps: its center is the mean, and its spread is the standard deviation. So the deciding question is not which curve looks familiar. What kind of data do you have? Are you counting notifications or other events, or measuring heights or another continuous value? Check real data before assuming a fit. For event counts in a fixed interval, choose Poisson; for continuous measurements around an average, choose Normal.