There are two reasons why people build analytics platforms:

1. to collect data that they can't get any other way
2. to run reports on data that's been collected

When you're thinking about the types of reports you can run on data, it seems that there are an endless number of ways to slice and dice.

There are.

The problem is that you should never ask an engineer for something so that you can query the data "any way we want". The result is always going to be messy and take a long time to build.

The answer is to start by producing one-off reports. Much the same as adding new data sources, the best way to manage this part of the product is to choose the 3-5 that you're going to use every single day. Start by building those in the first month, and then add more later. Remember the Lean mantra -- "I can always add it next month." 

One of the nice things about reporting tools is that you're not locked in by what you did last month. Even if there are some changes under the hood, it's going to be better for your company as a whole to start small, since you'd have to have done that anyways. It's better to spread out the development time so you have time to come up with better ideas once you've played around with the reports you've got now.