Startup Stack Vs Scale-Up Stack

About half the data stack that got you to 25 people won’t make it to 80. You can guess which half now.

What breaks first is the glue around the warehouse.

Scripts on one engineer’s laptop. Google Sheets as a source system. A BI tool where every dashboard has its own copy of the revenue logic. Permissions handed out by whoever had admin that day.

At 25 people all of that works, because four people know where everything is. At 80, finance, sales ops and a product analyst are asking questions at once, and nobody remembers why the churn query excludes March.

The parts that usually survive: a decent warehouse, dbt models if they were kept tidy, and a managed ingestion tool. Those were mostly built for the next size up anyway.

So the useful exercise now is labelling each piece: keep, replace at 50, replace at 80. Then the swaps land on your roadmap, on dates you picked, ideally before the quarter where you’re also hiring.

I get these calls wrong now and then. A tool I’d have bet on replacing stretches fine to 150. That’s the nice kind of surprise.

Which piece of your stack would you bet won’t make it to 80 people?