An objective names the change you want to make. Key results make progress inspectable. A useful key result has a measure, a baseline, a target and a time horizon. Targets in these exercises are fictional planning choices, not reported industry results.
An output measures work produced, such as screens launched. An outcome measures the effect, such as successful task completion. A leading indicator may change before the outcome, but its connection to that outcome is a hypothesis to check. A lagging indicator reports an effect after it occurs.
Add guardrails where improving one number could hurt another. Reducing handling time by rejecting valid requests would be a bad result. Define denominators and cohorts too: a conversion rate cannot be interpreted consistently if the population changes each week.
A metric needs a clear population and a check for side effects.