Katie Bauer!

The longer I spend in the data world, the more I'm triggered by the phrase "thought partner." It's always cited as an ideal role for data scientists and analysts to be in, but it's a slippery concept. Whenever I press, people struggle to say what specifically it means.

When product people say it, it often means something like "the magic person who uses Big Data™ to find a simple change I can make to give me a 10x advantage." This is obviously bad; maybe it could happen, but it probably won't. Classic expectations mismatch.

But what's annoying is that data people say it all the time too. For them, it often means something like "I want to make recommendations and have them acted on immediately." There's frequently an implicit assumption that PMs/engineers should seeking out data advice as well.

If this is your expectation, you will not like working in a data role that isn't pure machine learning/optimization. Data people are usually not decision makers, and they are not necessary to product in the same way that developers are.

PMs will not seek out your perspective unless they think it's adding something they don't already have. And no, it's not using Big Data™ to find a simple change that brings a 10x advantage. Making a PM value your perspective is far subtler and happens over a longer time horizon.

It's partly about being on their side, even when you think some ideas are dumb. Answer some seemingly pointless questions, help run experiments you think won't work. Don't go overboard, but be supportive. It builds trust, and it's damn hard to influence people who don't trust you