 We'll start by doing a quick overview of the attribution reporting API and how it works. So in step one, from a user's perspective as well as an ad tech's perspective, when a user views or clicks on an ad on a website, the ad tech can register this event, which is then stored in the user's browser. Then later on, if a user engages with the advertiser's website or converts on the advertiser's website, which could be a purchase or page visit or other types of activity, the ad tech can then go ahead and register this event as well, and this gets stored in the browser. The browser will then create a report by matching this conversion with the previous event of a user viewing or clicking on an ad, and then in combination with some privacy measures such as delays, noise and metadata limits, a report is delivered to the ad tech in a privacy preserving way. For the attribution reporting API, there are two types of output reports, summary reports, which provide detailed information in aggregate and event level reports, which provide limited information, but per conversion. For this presentation, Noise Lab focuses on summary reports and allows you to experiment with different parameters that may impact the amount of noise in your final summary reports.