Last week I had the pleasure of attending and speaking at the first in-person meeting of the New York Web Performance meetup since the pandemic started. In my eyes this meetup, organized by Sergey Chernishev and now Mellisa Ada, is an institution, being the first ever Web Performance meetup in the world, and so it was a thrill to be back with the tribe, so to speak.
After Sergey did what Sergey does great (give an intro to the meetup and its famous geekaways list useful sites and resources, encourage mingling), I proceeded to do my version of an Old Man Yelling at the Internet and complain about some of the performance studies circulating out there that I think are doing more harm than good. I wanted to make the point of not blindly trusting performance advice and numbers we read out there. Especially when said advice and numbers come without any data, methodology, sample sizes, and so on. I think it's a disservice to the web perf community to publicize poor studies. A situation I try to avoid is have a member of our community overpromise benefits to their business folks after reading a study and then not seeing the expected business metric impact from the performance improvements. And at the same time I'd like to encourage people to do their own studies and, if possible, do them out in the open where our peers can review and provide feedback.
The other part of my talk was showing how I did one actual study: gathering the data, experimenting, and looking at the results. The idea was revising an old study of progressive JPEG encoding that was almost irreversibly lost to the ether, but saved thanks to the Web Archive. The slides and code from my talk are available here. The quickest of TL;DRs: always run JPEGs through MozJPEG and for files over 2.5K it's likely that progressive encoding will produce a smaller file size.
Then came the really fun part of the evening when Scott Jehl demonstrated the new Opportunities and Experiments features of WebPageTest and played with a few sites, conducting live zero-code experiments. Did you know that the team even built a fake news website called The Metric Times and deliberately added anti-patterns to let anyone who wants to try out these zero-code experiments for free?
By far my favorite part was the Q&A session following Scott's demo. Lots of questions, suggestions, and thanks for the new features and for WebPageTest in general. It was touching to see people express their enthusiasm and support. Concerned questions like "I see thousands of tests queued at certain times at a test location, are you under attack?" reminded me of when I mentioned to Pat Meenan few years ago that I felt bad clicking "Capture video" as I didn't want to strain the WPT agent. ("Go ahead, capture video" is what he said, by the way).
I have been a WebPageTest user and a fan for a long time and it felt great to now be on the receiving end of the love (admittedly undeservedly, since I only joined the WPT team mere weeks ago).
The meetup was hosted by my new employer Catchpoint and it was the first time the New York WebPerf meetup group was coming back to meeting in person. We are hoping to host more in the future in other cities as well. Sergey mentioned how it's difficult to find venues to host meetups. It's always been hard, it turns out, but with many companies closing offices and transitioning to remote work, it's getting increasingly harder. In case you're wondering, vaccinations were required by the meetup and masks were highly encouraged and provided.
If you're interested in offering space for the NY WebPerf meetup specifically, contact the organizers Sergey or Mel and for any other meetup (web performance or otherwise) do get in touch with your favorite meetup's organizers.
Catchpoint is committed to continue supporting more WebPerf meetups and events in the future, so let us know if you'd like to attend a local meetup, host a meetup, offer space or connect with other meetup leads.
Photo credits: Kiera Ada
Stoyan Stefanov has recently joined WebPageTest by Catchpoint as an Software Engineer (formerly at Facebook and Yahoo!). He has a storied history as a very accomplished author, contributor, and custodian of the well known and respected Perf Planet calendar. He habitually speaks about web development topics at conferences and on his blog at phpied.com. Stoyan also created the smush.it image optimization tool and was the architect of Yahoo's performance optimization tool YSlow 2.0.@stoyanstefanov