Epistemic Status: Speculative
tl;dr: Impulsivity looks like it might be affected by how motivating competing actions are. If it's the total amount of motivation for other tasks, we should be able to remove weakly motivating distractions to increase productivity. If it's the average amount of motivation, we should be able to add many weakly motivating distractions to increase productivity. Which seems more effective?
The procrastination equation is supposed to describe one's motivation for doing a given a task. It goes like this:
motivation = (expectancy value) / (delay impulsivity)
Expectancy, value and delay all seem to refer to the task itself, but impulsivity seems to refer more to yourself or your context. It is a bit different from the other terms.
If we look at suggested ways to reduce impulsivity, we come up with things like turning off phones blocking distracting websites, or making failure more costly (perhaps by a commitment contract). Turning things off increases the delay for a distracting action. Increasing the cost of failure reduces the value of distracting action.
These methods of handling impulsivity point towards impulsivity being our motivation to do all the other things available to us.
What does this suggest for decreasing impulsivity and increasing motivation?
Well, if it is just the proportion of the (delay-discounted) expected value, then even if we have a very tempting alternative action, we should be able to increase our motivation just by eliminating the lower level distractions. Evidence for this is when you have choose a bad option over both a good option and an OK option because you are not motivated by the good option (example I stole from somewhere that I can't find again: you really want to go to bed and rest but you should do some homework first, and instead you end up browsing the internet).
If instead it is something like the motivatingness of our task relative to the average motivatingness, we could increase our motivation by adding additional unmotivating options to our environments (or, of course, eliminating the most distracting options). Evidence for this is the success of structured procrastination.
(This might connect to Bayesian brain-type stuff: just as a the probability of B given A is the proportion of the probability of B that comes from the probability of B and A, so it is tempting to look at our motivation to do a task as something like the proportion of (delay-discounted) expected value that comes from that task.)