Mandy Hanson, it is a great question to ask, although it is not so easy unfortunately. You can with some finessing get all this activity data out via the API for further analysis. But it requires some technical and BI expertise.
I think the more simple (but unfortunately time-consuming) method will be to create a lot of smart lists based on specific queries in order to see relative conversion rates of users who have done certain activities. Develop your hypotheses and then build smart lists to answer them. E.g.,
Total number of people have done x AND converted / Total number of people who have done x
= conversion rate associated with activity x
Keep doing this for different activities (watched a webinar, clicked an email, visited a pricing page, etc.) and you start to see which behaviours have a strong correlation with conversion (however you choose to define it).
If you want inspiration to develop your hypotheses you can always take a list of 100 converted leads and just manually review the activity history and look for patterns. Yes, you have to do this one at a time so it's not a technique for aggregate reporting, but it's a way of getting into the nitty-gritty of the data and getting your analytical juices flowing.
.png)