Transient photoluminescence allows us to quantify recombination by determining the characteristic decay time of an exponential decay. This decay time is frequently referred to as the charge-carrier lifetime and is often considered to be a single value for a certain perovskite film. However, we show that in lead-halide perovskites, the decay is primarily not following an exponential decay or a superposition of several exponential decays. This observation is somewhat counter to many previous findings on the topic, which we ascribe to the difficulties in distinguishing between multi-exponential and power-law decays if data is obtained with a limited dynamic range of 1 to 3 orders of magnitude (1). Here, I present a range of transient PL data obtained with larger dynamic range that thereby allows us to clearly distinguish decays that follow a power law from those that follow an exponential decay. Subsequently, I discuss the implications of power-law photoluminescence decays and, therefore, decay times that vary continuously as a function of time or injection level. The variation of the decay time by orders of magnitude implies that the decay time becomes a rather difficult to use figure of merit. One peculiar mathematical aspect of applying the concepts of exponential decays to power-law data is that the decay time follows the time after the pulse. Thus, if the time window of the measurement is increased (i.e. by reducing the repetition rate), the decay time taken towards the end of the decay also increases. As fitting biexponential functions to semilogarithmic plots of PL vs. linear time puts a strong emphasis on the end of the decay, the repetition rate also significantly influences exponential fits of power law decays. The peculiar finding here is that if decay time is presented without the associated carrier density at every point in the decay, it is meaningless as long as the decay is mathematically following a power law. A work-around for this problem is to either present the decay time as a function of carrier density or Fermi-level splitting or to directly determine the recombination coefficient of a bimolecular decay. I then explain possible physical interpretations of non-radiative bimolecular recombination coefficients and show how they are related to steady-state assays of recombination, such as the photoluminescence quantum efficiency.