We're at ~1/10th your scale but query speed is a major concern due to (rightfully) impatient clients when it comes to data viz. Unless you're using BQ tables as input for high throughput compute, I'd focus on optimizing your BI tool or creating analysis tables that would prevent end-user slowdown.
Ex: Recently created a big table (by materializing fact/dimension table joins and COALESCE operations) solely for analysis purposes. It sits "outside" our normal data warehouse setup mentally, but we can still maintain data quality/lineage as it exists inside dbt. Allowed us to do away with Tableau fixed calculations and cut load/group by times for end users ~95%.
Yeah, I think there is a tendency to over-simplify the duties of responsibilities of a MM as "printing money", as if it's a passive business that solely relies on existing infrastructure (of course, just the existence of a boat load of directors and MDs all with their niche electronic trading/MM products is an indication it's not so simple, but I digress) and scale -- that is just not true.
I try to explain to retail investors and outsiders interested in markets that this whole operation requires a truly mind-boggling amounts of money and brainpower to keep going on a regular basis, but it's hard talking to people who have never experienced working in the industry.
That's true to an extent at the base level, but there is a pretty large pre-trade/post-trade analytics infrastructure in place to analyze trade quality and market trends. It's not uncommon to lose weeks worth of profit due to predators dumping shares into slow-moving MMs since such predators can send orders faster than MMs can cancel them in reaction to significant market movements.
Some fundamental understanding of the asset class and market microstructures is needed (if not on the tech side then the quant side, but ideally both) so you know what you're trading. Given the sheer volume that goes through your typical broker-dealer, that's quite a reasonable investment to make.
Ex: Recently created a big table (by materializing fact/dimension table joins and COALESCE operations) solely for analysis purposes. It sits "outside" our normal data warehouse setup mentally, but we can still maintain data quality/lineage as it exists inside dbt. Allowed us to do away with Tableau fixed calculations and cut load/group by times for end users ~95%.