 Now we're ready to get started with our very first simulation scenario in NoseLab. First of all, I would like to show you the default dummy example measurement data we're using NoseLab. So here we assume you have an average daily conversion count per bucket over hundreds. You can also change that of course. And we assume that you're going to want to measure two metrics. The first one is purchase value and the second one is purchase count. You can edit all of that in advance mode. But this is kind of a boilerplate to get you started with your first simulation. We're also assuming that you want to track three dimensions. You want to track the campaign ID, the geography, and a product category. And we're assuming that you have four different campaigns, three different geographies, and two different product categories. So this makes for a total of four times three times two. So 24 buckets in total for each of the specific measurement goals. Right. These are the default parameters. I'm also using a batching frequency that is daily. We can also change that. We'll talk about that in a minute. And I have scaling on again one more parameter that we're going to discuss in more detail. You will also notice that here at the top of this parameter panel, we have contribution budget and epsilon. We will not be editing these for now. The contribution budget is an API parameter that's fixed at the moment. And as per epsilon, I'll touch on that in the next section. All right. So with this, we're ready to run our first simulation. So let's go ahead and run that. So here we have a little timestamp to remember when we run the simulation. And we can already look at our output for our measurement value, our average percentage error, which is the metric we're going to look at for now. RMSP will be useful later is about 1.8%. This is the exact value. But here knows that gives me a little overview that's colored. Give me an idea of how much noise that is. And for the purchase counts, I have an average percentage error of 0.1%. Important to note that noise is added from a random distribution. So if I was to run the exact same simulation with the exact same parameters, this percentage here, these may vary a little bit. They should stay in the same ballpark, but they may vary a little bit. This is the nature of noise. It is random. What you can see here is really the average percentage error that's calculated as an average on the whole summary report. The data table is over here. We've already touched on that in the introduction about the user interface. But this is really all of the output simulation data of NoseLab. Each key, this is the bucket. Over here, this column represents the summary value before noise, the summary value after noise, noise over here. You can see that noise can also be negative. This is important. And here, finally, we have our calculation for noise APE.