1. Define a green light criterion, and test a small launch against it. You have a new idea that you think will make your users happier or more productive? Grow your user base? Make the company more revenue? Great, build it and try it. Launch a small test or beta to a tiny percentage of your audience and have a criteria in mind for whether/when you should launch it to everyone. Is it if you get really positive feedback? X%+ more growth? Y%+ more revenue? Clarifying these criteria as early as possible (before launching, ideally) lets you be more objective when the rubber meets the road.
2. Define a sunset criterion. In some cases, it may not be practical or even possible to test your idea to a small group. If that’s the case, ask yourself this: Under what circumstances would it make sense to kill the feature or product after it’s launched? Again, it’s easier to be objective about this earlier rather than later.