 We've all experienced the frustration of trying to use a web page that just seems frozen and unresponsive to user input. And as bad as it is, this is a part of the user experience that we haven't really given much attention to in our analytics tools. In this episode, we're taking a closer look at a brand new metric called first input delay and how we can use it to measure these kinds of user experiences on the web. Let's get started. First input delay, or FID, is a new metric aimed at measuring that little bit of frustration we've all felt when we try to interact with a page. But nothing happens for a while. You may have already heard of a similar metric, Time to Interactive, or TTI. But the difference is that this metric only tells us how long it took for the page to become interactive, not the time to respond to a user interaction. The first time a user interacts with a page, like clicking, tapping, pressing the key, FID measures the time it takes for the browser to actually respond to that gesture. This also means that a page with no input gestures at all will have no FID. So don't count on it happening for everyone. Pages that do very little work on the browser's main thread should be interactive quickly enough to respond within reasonable limits. Studies show that a user will perceive a response time of 100 milliseconds or less to be instantaneous. So that's a good limit to stay under. But FID is not only affected by how much work the page does, it may also be affected by the users themselves. A page could have a really slow TTI, but as long as users are not trying to interact with it immediately, for example, if it has a captivating loading spinner, FID may still be fast. Or it may even be affected by users' hardware. A user on a low-end device with a slow CPU may experience longer TTI and worse FID. To help us understand just how frustrating the first interaction experience is on the web, let's turn to the Chrome UX report. As of July 2018, FID is available as an experimental metric in the data set. This means it's not quite standardized just yet, but we'd still like to publish the data for exploration. So let's write some queries to better understand the state of FID on the web. Accessing it is the same as any other Chrome UX report metric, only this field has an experimental namespace before it. Here, we're adding up the densities of all FID experiences less than 100 milliseconds for developers.google.com. And it turns out that this origin actually has more than 90% of so-called instant experiences. We can also query for the average percent of instant FID experiences across all origins. All we need to do is modify the last query to remove the origin filter and divide the total density by the number of origins in the data set. The global fast FID is 80-something percent, so not as high as developers.google.com, but still surprisingly good. How does the distribution look for FID beyond 100 milliseconds? We can add up the densities for all bins in all origins histograms to look at an average distribution. Visualizing the data in sheets, we can see that there's a spike of FID values around 225 milliseconds. This isn't slow, but to a user, this delay is definitely noticeable enough to feel somewhat unresponsive. From there, the long tail extends into the seconds. Cumulatively, almost 2% of FID values are one second or more. This is a brand new metric, so leave a comment below and let us know what you think of it. For more information, check out bit.ly-fid-docs. And of course, start exploring FID in the Chrome UX report and let us know what you find. Thanks for watching, and we'll see you next time.