How do you clean up 10 years of metric sprawl? Looking for a framework
Hey everyone,
I work for a company where metrics have never been properly governed. For the past 10 years, everyone has had direct access to the raw database, which led to a massive sprawl of metrics created independently by business, product, and data teams with zero consistency or shared standards.
I've been tasked with cleaning this up, and honestly I'm struggling to find a clear methodology to tackle it.
What I've figured out so far:
- Start by defining the core concepts ("base entities"): what counts as a user? What counts as a company? etc.
- Then map out the dimensions tied to those entities, for example:
- Active user → dimension
status: active / inactive - Companies by country → dimension
country
- Active user → dimension
My question:
What methodology or framework would you recommend for structuring this kind of work end-to-end? Where do you start, how do you prioritize, and how do you avoid drowning in 10 years of accumulated chaos?
Would love to hear from anyone who's been through something similar. Thanks!