I wonder how this could scale. Right now, the one data analyst in your post can handle everything; the iteration makes total sense.
How could this scale to a company where 1) there is a data/analytics engineer taking care of the semantic layer 2) a data analyst fetching the data 3) a business user who has input on the customer segments? I think figuring this out will be the biggest challenge. I'll keep thinking about it :)
There's definitely a subject of governance and roles here indeed. But does it need to scale?
I mean, yes the DE could approve/decline the new semantic definitions arising from the analys/business usage of that loop. But wouln't be a bottleneck? Or is data governance finally entering the loop?
I would rather move forward on the latter, i.e. business/analysts are fetching the semantic layer, asking questions not answerable -> that trigger the data governance operators (mixed team of DE and stakeholder) to approve/decline new semantic layer definition.
That way the semantic scale as more user enter the loop, but it's not exponential as the data governance "operators" are controling the process.
Hi Julien, this was super interesting!
I wonder how this could scale. Right now, the one data analyst in your post can handle everything; the iteration makes total sense.
How could this scale to a company where 1) there is a data/analytics engineer taking care of the semantic layer 2) a data analyst fetching the data 3) a business user who has input on the customer segments? I think figuring this out will be the biggest challenge. I'll keep thinking about it :)
There's definitely a subject of governance and roles here indeed. But does it need to scale?
I mean, yes the DE could approve/decline the new semantic definitions arising from the analys/business usage of that loop. But wouln't be a bottleneck? Or is data governance finally entering the loop?
I would rather move forward on the latter, i.e. business/analysts are fetching the semantic layer, asking questions not answerable -> that trigger the data governance operators (mixed team of DE and stakeholder) to approve/decline new semantic layer definition.
That way the semantic scale as more user enter the loop, but it's not exponential as the data governance "operators" are controling the process.
cc https://substack.com/@thedatagovernanceplaybook