Mark Zeman
@markzeman.com
980 followers 67 following 20 posts
Founder @speedcurve.com, Hopeful idealist
Posts Media Videos Starter Packs
Pinned
markzeman.com
Loving the shift away from bloated client-side JavaScript. With RUM tools offering better visibility into the impact of JS, the DX vs. UX trade-off is becoming harder to ignore. Any UX focused changes in your stack making it to production? Did your numbers improve? www.speedcurve.com/blog/long-an...
Reposted by Mark Zeman
tammyeverts.com
"How fast is fast enough?"

If you've ever asked this question, then you might want to join tomorrow's SPDY STREAM!

Henri and I will be chatting about how to answer that question, among other things. See you there at 10am PDT / 1pm EDT!
markzeman.com
Loving the shift away from bloated client-side JavaScript. With RUM tools offering better visibility into the impact of JS, the DX vs. UX trade-off is becoming harder to ignore. Any UX focused changes in your stack making it to production? Did your numbers improve? www.speedcurve.com/blog/long-an...
markzeman.com
I want all my web performance metrics described with baking analogies. Who knew that cake had "Frontend" time. Although that would mean that a cake's Time To First Bite (Backend) would come after Frontend!
Reposted by Mark Zeman
tammyeverts.com
Ever wondered why your business and engagement metrics didn't improve after you sped up your pages? You need correlation charts.

👍 Validate your #webperf & #UX metrics
👍 Identify your site's "performance plateau"
👍 Spot performance-blocking trends on your pages

www.speedcurve.com/blog/site-sp...
SpeedCurve | Correlation charts: Connect the dots between site speed and business success
If you could measure the impact of site speed on your business, how valuable would that be for you?
www.speedcurve.com
markzeman.com
There is also a 4th, slightly more controversial optimisation trick for SPAs as well.
speedcurve.com
Single-page apps have unique #webperf challenges. Here are some common performance issues with SPAs and three things you can (and should) do to optimize for them.

www.speedcurve.com/web-performa... #ux #pagespeed #sitespeed
Optimizing Single-Page Applications (SPAs) | SpeedCurve
Here are some common web performance issues with SPAs, and how to optimize them.
www.speedcurve.com
markzeman.com
Love them or not, you’ve got to be up to speed on Core Web Vitals.
speedcurve.com
Need to get up to speed on Core Web Vitals? This guide is a great place to start!

👉 What are Core Web Vitals?
👉 SEO impact
👉 Business impact
👉 How to investigate and fix common issues with LCP, INP & CLS

www.speedcurve.com/web-performa... #corewebvitals #webperf #ux
Get Started with Core Web Vitals | SpeedCurve
Everything you need to know about Core Web Vitals – from SEO and business impact to how to monitor and fix issues with each Vital.
www.speedcurve.com
Reposted by Mark Zeman
markzeman.com
How often do you find yourself in this position though? Where LCP is not actually what most people would pick as LCP when looking at a page. When LCP is unintuitive and complex it breeds disinterest.
Reposted by Mark Zeman
tammyeverts.com
I just had a good chat with a @speedcurve.com customer who wanted to know why their CrUX numbers didn't match their RUM numbers. It was a good reminder that even experienced #webperf / #ux folks might not realize that CrUX isn't full RUM.

Helpful explainer: support.speedcurve.com/docs/speedcu...
A graphic showing the following list of bulleted text:
CrUX is not the same as RUM.
CrUX is only collected from Chrome browsers, while RUM looks gathers data from all browsers. (This is a huge factor.)
Sample sizes are probably different.
Data aggregation is probably different.
CrUX data is rolled up monthly, while your RUM time period can vary depending on how you're looking at your data.
You may be comparing substantially different dimensions/cohorts
CrUX only segments a site by origin, while RUM allows for deeper segmentation.
markzeman.com
A well articulated debug Harry, but it distresses me that this is where we at as a perf community. Has LCP really made it easier for devs to improve user experience? The detailed caveats and nuances you had to go through to understand LCP are beyond what we should consider a “good” metric.
markzeman.com
Around 12-15 days. We’re taking it pretty easy so aiming for 15 days. 100km per day average.
markzeman.com
It’s an antique now! Should really be on display somewhere.
markzeman.com
Super excited to see this bike packing build come together. Will be riding the Sounds to Sounds in March. www.touraotearoa.nz/p/sound-to-s...
Half built Viral Wanderer bike.
markzeman.com
As a dyslexic, not being able to edit posts on BlueSky is a bit rough. 😢
Reposted by Mark Zeman
nucliweb.net
I love my new @speedcurve.com #perfhero sticker 🥰
Photo of part of my cover laptop with stickers
markzeman.com
Web font subsetting with Unicode ranges is such an awesome and under appreciated feature.
stoyan.me
Say hello to `inverse-subset`! The idea is from a conversation with @andydavies.me at #perfnow. When you have a complete font and a subset e.g. Latin, generate another subset consisting of everything not in the original subset. E.g. all characters that are NOT Latin. npmjs.com/package/inverse-subset
inverse-subset
A CLI tool to generate inverse subset fonts using a complete font and a subset font.. Latest version: 1.0.1, last published: 4 minutes ago. Start using inverse-subset in your project by running `npm i...
www.npmjs.com
markzeman.com
We done this on SpeedCurve for ages. Works great. Unicode ranges don’t get enough attention. Super powerful feature.
markzeman.com
Amazing to see Element/Container Timing gathering steam. These are fundamental for truly connecting web performance and user experience. Huge thank you to @jasew.bsky.social and @yoav.ws for pushing these along.
webperfwg.org
In our meeting yesterday, we decided to adopt Element Timing (and the followup Container Timing work) into the working group. CfC coming up!!

We also discussed some upcoming improvements to the crash reporting API (still in incubation).

Minutes: w3c.github.io/web-performa...
w3c.github.io
markzeman.com
CWV would be great as iOS is a huge blindspot for a lot of people and is a significant amount of traffic. Even better would be element/container timing. Apple values great user experience yet there’s very little in WebKit to help us measure the performance of what user actually see.
Reposted by Mark Zeman
nicolesullivan.bsky.social
Would you want core web vitals In Safari? Can you help me understand how you would use them and why numbers in one browser isn’t enough?

(These may sound like silly questions, but I’d love to understand *your* specific context and use cases)
Reposted by Mark Zeman