I have completed an experiment to measure the decay rate of an isotope, and I am trying to estimate uncertainties. The half life is 40 seconds, with decays counted over a 15 second period (with many of these 10 second periods for a total of 6 minutes of recordings) However in more detailed readings I think my set up may not model a Poisson distribution. As the rate of decay is clearly will drop off significantly over the measurement timeframe. So am I measuring a non-homogenous Poisson process or homogenous, or am I completely missing the understanding of non-homogenous Poisson? Any advice would be greatly appreciated, my confusions are doing my head in.