Impatience, Laziness, Product
Have no time for bullshit rabbit holes that your users don't give a fuck about. Build for your users and your company's success, not fringe ideas that only one person can explain.
If it takes an analyst a month to prove that an idea about changing button size might work to increase sign-ups, have no patience for that. If you have to back in to the "why," then fuck it. What's the more obvious idea that will give you big momentum? You should only have patience for mission-critical expeditions that slap you in the goddamn face. Look into the obvious stuff immediately, and archive the little shit until your big metrics are healthy. Constantly be clocking suggestions and ideas, but always be looking for big, shiny, sexy things (impact).
What the hell is it?
Constantly say "yes, and..." so that you don't dismiss something cool, but quickly ingest and gut-check the idea. Get really lazy about it immediately: how obvious is this idea? Why? If it's going to take you three weeks to prove as a valuable idea, ugh. Why? Use your time wisely on this, and only if you think it's a real headliner. Don't work on lame stuff that won't change the game - does your team really have nothing better to do?
This means you need to be lazy and high-level about concepts, even once they're sticking. Always tie them back to the big metrics when you're vetting an idea, since that's really all that matters. If you can't, then fuck that idea for now, and assume you haven't earned the right to work on small optimizations yet.
Have a bit of product ADD and get distracted by better stuff. Then do just enough to prove success or failure and move on.
Are you sure?
If you've got something that's screaming "do me," then question it and all its parts again with engineering. Why are we doing this? Seems like a fuck-ton of work. Instill a bit of doomsday in yourself. Ask of yourself and your team, if you could only build one thing for this company and it's our last day on Earth, would it be this?
Then, be a little under-thorough and non-committal in designing your product at first, too. De-scope everything that's too detail-oriented or time-consuming if it won't directly help you prove this thing is good. Assume throwaway work and total failure.
Ok, build It.
If you think you trust an idea enough to work on it, get it done sooner, as soon as possible, now, yesterday. But get it done just enough to learn if the idea really was any good. Always ask, what's taking so long? Be impatient when determining scope and process. Can you wait for results? If you can, then is it really your top priority? If the results are not stupid obvious with only light analysis, err on the side of failure and proceed with caution if you don't start over.
Be budget-minded so you can execute as many ideas as fast as you can, hardly building each one. Laziness on each project, but hyper-activity on the whole.
Stop being so lazy.
Only when you nail the concept should your team kick into OCD high-gear to button-up all the edge cases and squash all the bugs. Only then do you demand rigor and strive for detailed excellence. Because you should always be about that for the shit you believe in and work on. Now you can move beyond the must-have features and spruce it up. A mistake in font weight will not affect bottomline conversion rate, but does make you look sloppy as fuck if you're going to keep the feature up for more iterations.