While I agree that using percentages would make impact more comparable between agents and timesteps, it also leads to counterintuitive results (at least counterintuitive to me)

Consider a sequence of utilities at times 0, 1, 2 with $U_{0} = 1$ , $U_{1} = 0.01$ and $U_{2} = 0$ .

Now the drop from $U_{1}$ to $U_{2}$ would be more dramatic (decrease by 100%) compared to the drop from $U_{0}$ to $U_{1}$ (decrease by 99%) if we were using percentages. But I think the agent should 'care more' about the larger drop in absolute utility (i.e. spend more resources to prevent it from happening) and I suppose we might want to let impact correspond to something like 'how much we care about this event happening'.