[hpaavola]

If the oobjective itself is easy to measure, "Increase purchases through Alexa’s voice interface" for example, then that is the the key result. Objectives need to be abstract complex things, grandiose goals. And key results are measurable good ideas to approximate our movement towards the objective.

[markerz]

I don’t think objectives need to be complex or grandiose. Sometimes they can be but rarely do they need to be. In the end of the day, engineers are just employees solving business problems. The main business problem is driving shareholder value. Everything else is a means to achieve that. Fetishizing complexity leads to unnecessarily complex systems to achieve our personal objective of promotions and raises.

At my previous company, there was a strong drive to make OKR’s complex because it looks better on a performance review package, but it just caused us to spend so much time creating and finding complex problems rather than solving the ones in front of us.

[phillipcarter]

IMO "Increase purchases through Alexa’s voice interface" is pretty grandiose at the scale of Alexa, given that it's a huge network of devices and services powered entirely by voice.

[asplake]

“Grandiose”? Stretching, sure, but for a time-bound objective, oh please no.

What leading indicators would suggest that you’re making real progress towards that goal? Those would be key results, more granular than that objective. And how might you bring them about? That they inspire solution ideas is the point really.