So, again, the use cases here are varied. For the most common, which is understanding user behavior, you'll do something like: - I'm going to capture 5% of us

So, again, the use cases here are varied. For the most common, which is understanding user behavior, you'll do something like: - I'm going to capture 5% of users sessions as they progress through a purchase funnel - I capture data for a few days after a design change, and look at behavioral changes, such as spending more time in a particular step of the funnel - Where there's something weird going on, I'll watch an aggregate view of people's mouse behaviors - I can then see that a lot of people are hovering between two different buttons, or highlighting a particular sentence - Based on that feedback, I'll do an A/B test of changing the button textx or the dialogue message - Capture more data, did that change? It's not "Let's go look at john smith's website behavior" except in the support example, where you get a user's consent to do that, and you usually only do it for internal corporate applications where there's a dedicated helpdesk.