What does "good" look like for a data science org? And how do you get there?

For all the managers / team leads out there, how do you know when your team is performing at a high level? For context, I work as a product data scientist and much of our work is evaluating features either through experiments or causal inference (not the typical ML Engineer role). And while I feel like we do a reasonable job, I'm never quite sure how an "excellent" team would approach the same work. Are there benchmarks or heuristics that …

causal context data data science datascience data scientist features good leads look managers product science team through work you