Interested in a new permanent or contract developer job? We can help find one.

Speed Perception and Lighthouse with Estelle Weyl and Parvez Ahammad

June 23, 17:00

San Francisco, CA
This location is shown only to members, 345 Spear Street, San Francisco, CA

External Registration

Open Registration Page

Speed Perception and Lighthouse with Estelle Weyl and Parvez Ahammad
No one likes slow loading webpages, but what aspects of page loading contribute to user perception of “speed” and how can we measure it?
Since people primarily consume the web visually, the visual perception of the webpage loading process should be of primary importance. Traditional web performance metrics defined in W3C standards are objective metrics, focusing on timing each process along the content delivery pipeline, such as Time to First Byte (TTFB) and Page Load Time. What metrics are most correlated with how human end-users perceive the webpage loading process: the above-the-fold rendering in particular? We’ll discuss how we  can try to objectively measure something as subjective as the  notion of “feels fast”.
Speed Perception: Measuring Perceived Performance
Sites need to feel like they’re fast: perceived performance is key. Making it feel 25% faster is a valid goal, but perception is subjective and we need to demonstrate subjective improvements with objective measurements. Oddly, studies show perf metrics don’t provide much insight into perceived performance. In this talk, we’ll look at what page load aspects affect speed perception, introduce the Perceptual Speed Index (PSI) which is a component of Lighthouse, and discuss the results of the first SpeedPerception.com study which have led to a combo-metric that models speed perception, which we are currently studying in phase 2.
Clearly, no one likes slow loading webpages, but what aspects of page loading contribute to human end-user perception of “speed”? “SpeedPerception” is an open-source experimental framework and a study that we developed to understand what “slow” and “fast” mean to the human end-user.
Traditional web performance metrics defined in W3C standards focus on timing each process along the content delivery pipeline, such as Time to First Byte (TTFB) and Page Load Time. We wanted to tackle the web performance measurement challenge by looking at it from a different angle: one which puts user experience into focus. Since people primarily consume the web visually, we focused on the visual perception of the webpage loading process. SpeedPerception is a systematic study of how human end-users perceive the webpage loading process: the above-the-fold rendering in particular: objectively measuring the subjective notion of “feels fast”.
Studies have shown that instead of only focusing on page load time, if we shift the focus to address perceived performance directly, we can make our web applications feel even faster even if we have to download the same number of bytes. What metrics do we need to measure to objectively test if a site load feels fast? We’re looking into the impact of blanks screens, long DOMContentLoaded events, pop ups, passive versus active wait times, jitter, and other factors have on perceived performance, and how to measure the improvements in these features. We looked at browser metrics like time to first byte, render start, visual completion, etc., to construct a fusion model that can give the most insight into subjective judgments of “Speed Perception”.
Speaker: Estelle Weyl

Estelle Weyl (@estellevw) started her professional life in architecture and then managed teen health programs. In 2000, Estelle took the natural step of becoming a web standardista. She is the Open Web and performance evangelist for Instart Logic and has consulted for Kodak Gallery, SurveyMonkey, Samsung, Yahoo, Visa, and Apple, among others. She is a coauthor of Mobile HTML5, CSS3: The Definitive Guide, and HTML5 and CSS3 for the Real World. While not coding, Estelle works in construction, de-hippifying her 1960s throwback abode.
Speaker: Parvez Ahammad

Parvez Ahammad (@perceptPA) leads the data science and machine learning team at Instart Logic. His group is focused on creating data-driven algorithms, and innovative product features that optimize and secure web application delivery at scale. Parvez has a PhD in Electrical Engineering and Computer Sciences from UC Berkeley, with an emphasis in computer vision and machine learning. When he’s not busy designing algorithms to mimic human perception, he enjoys being a dad to his two young kids and perfecting new ways to make them laugh.
Event details
Note: Price: $10 online, $15 at the door. No refunds. No ticket transfers.
We charge this small fee to help reduce no-shows. Please contact the organizers (sfhtml5 at gmail dot com) if this is a financial hardship for you. 
Schedule
5:00 PM Doors open, eat and drink (food and drinks catered by Google; limited veg/GF options available; special cocktail by Alcademics; wine tasting by Grant Marie Winery. 
Note: The event is on the 7th floor.
6:30-6:40 PM Announcements 
6:40-8:00 Speed Perception: Measuring Perceived Performance" with Estelle Weyl and Parvez Ahammad
8:00-8:30 Break
8:30-8:35 Lightning Talks*
8:35-9:00 PM Q&A with both speakers 
9:00 PM Raffle -- have a chance to win a Google Pixel phone (must be present to win)!

9:30 PM Room cleared
* Contact the Organizers if you'd like to give a lightning talk. 
** You must be present at the venue to qualify for raffle prizes.
Notes
* There is bicycle parking in the garage below the building, but you must bring your own lock and leave your bicycle at your own risk.  * There is also street and garage parking. Check all signs for restrictions and fees.  * Smile! We are recording and photographing this event. If you do not wish to be included in SFHTML5 video or audio recordings or photographs, you can watch the recordings of all of our events at bit.ly/sfhtml5videos from the comfort of your home.
Cocktails by Alcademics

Wine Tasting by Grant Marie Winery

SFHTML5 SFHTML5

Propose talk to SFHTML5