One strategy is to consider establishing a success metric as the first step to every project. If you can't design a measurement for success/failure, maybe you shouldn't be doing that project.
This is the one point that I keep re-iterating to the teams I work with. If you can't define what a success is move on and do something else where you can. It's very easy to get caught up in interesting stuff 'just because' but if it does not contribute in a measurable way then it's just another distraction that you should avoid.
