In a seed round, you need to both find PMF *and* grow to a few thousand users. This model helps you plan that growth.

Experimentation is easy in theory and hard in practice. Many effects can cause us to draw incorrect conclusions. Some...

Analytics is hard, but it often fails because it's not communicated efficiently by analysts to business people.

Most of the stuff we measure should go up and to the right. But what about stuff where the goal is 'this never happens'?

When you have a few data points, and no model of the distribution they are drawn from, you can still use what you have.

Understand, Identify, Execute is at the heart of Facebook's product management framework and key to its success.

A mental model of blockers can help understand why we need a regime of experiments rather than just one to fix an issue

There's no tradeoff between qualitative and quantitative analysis. Do both. The real tradeoff is something else.

I randomly discovered a history of interest in data and visualization running through my ancestry, dating back to 1944

There's a range of options for "deliverables" from applied researchers in the industry. Understand their tradeoffs.