this is excellent. I do very similar work to the youtube example and when i am not too lazy to write out a metrics tree / hypothesis tree, i find it extremely helpful. Good motivation to explicitly use framerworks
A fantastic guide, very well articulated and practical! I've been looking for something like it to make a handbook for my data team. Problem solved. :) Would love to hear your thoughts on mastering storytelling.
One important thing to consider during the defining & scoping phase is whether the metric change is meaningful/significant enough to warrant investigation. E.g. Does a 10% drop in total views last week fall outside the range of trivial fluctuations for the metric? We have been adopting a process control framework to add some rigor to our intuition. Here's a wonderful essay from Cedrick Chin if you are curious to dive deeper - https://commoncog.com/becoming-data-driven-first-principles/
this is excellent. I do very similar work to the youtube example and when i am not too lazy to write out a metrics tree / hypothesis tree, i find it extremely helpful. Good motivation to explicitly use framerworks
Glad you enjoyed the read.
It’s always good to use some of kind of structure to guide the workflow for sure :)
A fantastic guide, very well articulated and practical! I've been looking for something like it to make a handbook for my data team. Problem solved. :) Would love to hear your thoughts on mastering storytelling.
One important thing to consider during the defining & scoping phase is whether the metric change is meaningful/significant enough to warrant investigation. E.g. Does a 10% drop in total views last week fall outside the range of trivial fluctuations for the metric? We have been adopting a process control framework to add some rigor to our intuition. Here's a wonderful essay from Cedrick Chin if you are curious to dive deeper - https://commoncog.com/becoming-data-driven-first-principles/