Track your estimate accuracy over time. If you are consistently 30% over, build that into your buffer. Data-driven estimates beat optimistic guesses.
Strong answers include: breaking work into small tasks, using historical data and velocity, adding buffers for unknowns, distinguishing between effort and duration, involving the team in estimation, and tracking estimate accuracy over time. When estimates are wrong: communicate early, assess impact, propose options, and learn for future estimates.
Core PM skill. Candidates who always estimate optimistically are dangerous. Those who always pad excessively are wasteful. Look for evidence-based estimation and honest communication when things go wrong.