 I'm going to be really quick on that one and not really demo a whole lot of that. Why? Because it's documented in detail in our User Guide for a NOS Lab as well as in our Developer Documentation for Summer Reports. Let me just quickly introduce the concept. So far, when it comes to configuring buckets, we've only talked about the number of dimensions you can have and the number of possible different values for this dimension as well as the measurement goals. So all of that is something you can configure flexibly in NOS Lab, but in reality, you could do more. For a given set of dimensions and a given set of size for each of these dimensions, you could decide to configure various aggregation keys. So you can decide to configure your buckets a little bit more flexibly and the way you configure your buckets is going to have an impact on noise. And this is one more strategy for you as an API user if you'd like to reduce noise or to adopt a little bit more of an advanced strategy, you can leverage different key structures in order to achieve that. So in NOS Lab, we here have a setting for you to generate your key structure. You can define the number of key structure you want and you can define which dimension you would like to have in each key structure. So that's a more advanced way to configure your buckets. If you're curious about the type of impact this can have on noise, check out our user guides and check out our NOS Lab and how they simulate that.