While accurate, this setup did not scale well and was not easily automatable. Our first investigation in this area used HDMI capture hardware to capture visual data, ensuring we had no external factors affecting performance. We needed to measure additional metrics that were more representative of the visual user experience. Although there was a performance deficit, actual performance as perceived by the user wasn’t as dramatic as the 2-3x differences we observed in when the load event fired. To make our page load times more competitive it was important to understand how we compared to Chrome in practice from a user experience point of view. We set an aggressive goal to bring our page load performance to within 20% of Chrome on a platform-specific top list of sites on Desktop & GeckoView in 2019. The data confirmed that we had a significant deficit for a metric our users care about.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |