Interesting point! But something to note: this stack isn't actually that expensive - BigQuery in particular is free/cheap at low volumes, Airbyte + dbt are open source (and you can even host dbt for free with a free tier CI/CD tool like Github actions), and Hyperquery is quite cheap as well (it'll be cheaper and more stable than hosting open-source BI tools).

I think the two parts of the stack that'll really eat up cost eventually are the warehouse and ETL. There could be a way to consolidate these using an HLTP db. Or if you really want to reduce costs you can just look at data where it is, directly querying postgres, using mixpanel/amplitude's free tier for event tracking analysis, or manipulating data in Google sheets. The question is whether you believe the value of having more robust analytics as a start-up is worth the expense, and if it's not, then just stick with an even more minimal cost stack. 🙂