I’ve finished converting Fishtown’s Stripe package to work correctly (mostly?) with BigQuery as our target for the Warehouse. With no real documentation, that was a PITA, but now I’m trying to use the resultant data and I’ve got questions… sooo many questions.
While it looks like it’s (mostly?) doing what it’s supposed to, I’m seeing some odd results that maybe a minimal set of documentation or examples would help to clarify. Is there any documentation around this package and its models?
Hey @warpraptor! This is a totally fair question. The very short answer is that Fishtown’s data-source-specific packages are things that we’ve built up as a part of our consulting business over the years and are primarily intended to be used for that purpose. We opened them up to the community mostly as examples of what could be done in dbt and to get people started in the right direction.
Specifically on Stripe, the initial versions of this package were built over three years ago and we’re no longer actively maintaining it or using it on our own projects. It does some interesting / useful things in terms of modeling daily MRR data, but we also found that we weren’t really able to generalize MRR modeling in the way that we wanted to; in every client project we implemented, customizations on top of this baseline required us to override a lot of the stuff that the package was providing, so much so that we no longer really continued to find it useful.
I hope that you’ve found that the package provided you with some helpful tools when thinking through your MRR modeling! It’s a non-trivial business domain to model effectively.
Totally fair and understandable. Thanks. I’ve already started my own (maybe less accurate) but easier to calculate and more importantly easier to explain to non-tech, non-data folks version of the MRR calcs. I was looking less for a blow-by-blow walkthrough of the code/logic than a 1-paragraph summary of what the that final stripe_mrr table was showing me (explanation of the one-month delay, which of those final views would let me project the monthly MRR on a dailyi basis, etc).