Hi <<First Name>>,
Wow... that's all I can say about the volume of amazing new performance and observability research and resources that I've seen over the past month. There's so much energy in our community right now!
In an attempt to squeeze more into this newsletter, I'm changing up the format. There's so much good stuff! Keep scrolling and reading – and let me know if you like it. And please send me more links!
Until next month,
Tammy
@tammyeverts.com
|
|
|
The fastest and slowest US travel sites
It's spring break in many parts of the world, which means a lot of folks are travelling right now... which means a lot of folks are accessing travel sites via their phones over unreliable connections. So let's take a look at our Industry Benchmarks dashboard and see how leading travel sites perform on slower mobile devices.
Despite the limitation of a slow connection, it was great to see a couple of sites — RentalCars.com and TripAdvisor – deliver Largest Contentful Paint (LCP) times under 3 seconds.
At the other end of the spectrum, some sites had LCP times of up to 34 seconds. :(
What do the slower sites have in common:
- Slow backend times
- LCP element is delayed (in one case, it was the 131st resource on the page!)
- Too many resources in the critical rendering path (optimizing the critical rendering path is the key to delivering an experience that feels fast)
*Industry Benchmarks is a public-facing, interactive set of dashboards. Every day we test and rank industry-leading websites based on how fast their pages appear to load from a user’s perspective. It's a handy tool for benchmarking your own site, as well as digging down to see how the fastest and slowest pages in the dashboards are built.
|
|
|
Research
The State of Global Website Performance in China — If you're responsible for a website with China traffic, take note. The team at Chinafy analyzed how 614 enterprise websites actually perform inside China. Among many other things they found the median time for perceived rendering was 17.2 seconds and 66.4% of global sites did not load successfully.
JODIE CHAN (CHINAFY)
The speed of trust: How slow pages hurt your brand — Pinning down the impact of site speed on ephemeral things like brand perception is tricky. We didn't set out to do that in the research project I cite in this post, but it happened anyway!
TAMMY EVERTS (SPEEDCURVE)
|
|
|
Case studies
Investigating a High LCP for Nike.com — Aaron analyzed Nike.com’s home page to figure out why its Largest Contentful Paint (LCP) is much slower than it should be... and then showed how it could be sped up by 3.6 seconds!
AARON T. GROGG
How Subito Reduced INP by 100ms+ — Francesca shared how Subito dramatically improved INP by focusing on dropping under-performing third-party tags, improving their telemetry, and enabling the React Compiler.
FRANCESCA MILAN (SUBITO)
Cutting Load Times by 75% at Lorikeet — Three well-known optimizations (HTTP compression, asset preloading, and tree-shaking) cut web app transfer by 75% and chat widget time-to-ready down to 681ms. None of these tactics were novel, but the compounding was.
RYAN TAN (LORIKEET)
|
|
|
The Bigger Picture
The Main Thread Is Not Yours — When a user visits your site or app, their browser dedicates a single thread to running your JavaScript, handling their interactions, and painting what they see. As Den states, when you internalize that the main thread belongs to the user, not you, you realize that blocking the main thread starts to feel like taking something that isn’t yours.
DEN ODELL
Performance is UX: What Web Performance Optimization Can Do for UX Researchers — There's a natural alignment between web performance and UX, but we're often separated by different processes and a wall of jargon. Ethan offers great insights and recommendations for coming together.
ETHAN GARDNER
|
|
|
Getting Started with Native RUM and OpenTelemetry
You're already capturing how your site performs. This session is about extending that visibility to mobile, where unreliable networks, device fragmentation, and missing session context create a different class of problem.
In this 45-minute session, you'll learn how mobile teams use OpenTelemetry to capture real user telemetry, investigate crashes and performance issues, and integrate mobile signals into your existing observability stack — including a live demo from instrumentation to root cause!
|
|
|
Best Practices
Practical guide to the <img> element: From the basics to LCP — A hands-on guide to using the HTML <img> tag properly so your images load fast, look right on any device, and don't mess up your site’s performance.
JOAN LÉON
JPEG Compression — A super-clear, wonderfully nerdy walkthrough of how JPEG actually works under the hood, showing how it shrinks images by tossing out the stuff your eyes won't really notice.
SOPHIE WANG
Fast Video Banners That Work on Every Device — A practical guide to making video banners load fast and look right on every device by faking a “responsive poster image” so users see something nice instantly instead of a slow, janky video.
ESTELA FRANCO
TTFB Doesn’t Mean What You Think It Means — Reality check: TTFB isn’t one clear metric. It's a messy bundle of different timings that everyone defines differently, so comparing it across tools can be seriously misleading.
ROBIN MARX
Serving Static Content with Cloud Storage? Don't Forget the CDN! — Important reminder that dumping your static files straight out of cloud storage might seem fine, but without a CDN in front you’re quietly making your site slower than it needs to be.
PAUL CALVANO
Compression dictionaries — Ryan invited web performance heavyweights Pat and Yoav on his Lessons of a CTO podcast for a deep-dive chat about how compression dictionaries work, why they can beat gzip/Brotli in the right cases, and what it actually takes (and what can go wrong) when you try to use them in the real world.
RYAN TOWNSEND, PAT MEENAN, YOAV WEISS
|
|
|
|
