Hi, I'm Arjen. I fix Core Web Vitals.
Call me when you've hit a wall. I find what's actually slowing your site down, fix it with your developers, and leave your team with the monitoring and skills to keep it fast.
Right fixes, wrong order
You've worked through the ticket list, maybe for months, and the site is still slow. I see this on most projects. Every ticket makes sense by itself, but page speed is an ecosystem: the fixes depend on each other. A preloaded hero image gains you nothing while 500ms of scripts block the main thread in front of it.
So I start by finding out what has to happen first on your site, and we work down from there.
I start with your real visitors
Every job starts with field data: your p75 on mobile, which pages fail, and whether it happened overnight or crept up over months. I use CoreDash, the RUM tool I built, to see which element or script causes the delay. That's what we fix first, because that's what your visitors wait for.
You decide who builds it
Your own developers with my plan, me writing the code myself, or me working alongside your team for a few months. Stuck on one question? Then an hour is often enough. See the services
It stays fast after I leave
A fix is only done when your real visitors notice it. CoreDash keeps watching after the work is live, so a new plugin, app or deploy that slows things down gets caught the day it happens.

Loop Earplugs
"Mobile got about 800ms faster and checkout conversion went up 7%. It paid for itself pretty much straight away."
Nathan Ziarczyk
Results

15 marketplaces, featured on web.dev
I reorganized how third-party scripts load across 15 Adevinta marketplaces. Google published it as a web.dev case study.
Fixed in one session
Their team had been working on layout shifts and LCP for months. The cause was the product gallery, built entirely in JavaScript. We made the images render in plain HTML first and let the script add the scroller and zoom afterwards. The field data showed the result within a week.
A new plan every quarter
Since 2024 I've sat down with Prijsvrij's team at least four times a year. We look at the field data together, decide what matters most for the next few months, and work out how to build it. Their developers do the work. I make sure it's the right work, in the right order.
What clients say
Rituals
"We used to break performance every other sprint. Arjen put performance budgets in our pipeline and we haven't had a regression since."
Laura Vogels
Miro
"Our dashboards were doing way too much work on the main thread. CPU usage went down 25%, and our users noticed it right away."
Suzanne Beek
Swift
"The layout shift on our checkout is gone. CLS went from 0.4 to 0.02, on mobile and on desktop."
Louis Breton
DPG Media
"He found problems in our component library that we had been living with for two years. Within a few days you could see the difference."
Ashkan Malekloo
Zalando
"We've had audits before, and they all gave us a long list of problems. This one told us what to fix first, and why."
Alza
"Arjen taught our team how he works. We now find and fix performance problems ourselves."
Voices.com
"Arjen played a key role in fixing our Core Web Vitals. He got our critical profile pages back into the green in PageSpeed. What I value most is his hands-on attitude: he doesn't just find the bottlenecks, he works directly with our development team until the fixes are actually live."
Jayme Cousins
The tool behind the work
CoreDash: Real User Monitoring for Core Web Vitals
I built CoreDash because I needed it. On every project I have to see, while it's happening, which element or script is making a page slow or a click lag. After years of digging that out by hand, I turned it into a tool. It stays running when I leave, and it has saved more than a few of my clients' necks.
Real time, down to the cause
See what your visitors experience right now, and the exact element or script behind every slow page.
Works with your AI
Ask your AI assistant why checkout got slower yesterday. CoreDash answers through MCP.
Data stays in the EU. No cookies, no IP addresses, no personal data, so no consent banner. And no limits on how many pages you track or how long you keep the data.
Hit a wall? Tell me about it.
Send me a message with your URL and what's going on. I'll take a look and tell you what I think is wrong and what it would take to fix it. And if it turns out you don't need me, I'll tell you that too.
Mobile LCP is stuck around 3.8s. We went through the whole Lighthouse list and saw no improvement.
https://www.example.com/Had a look. Your A/B testing tool hides the entire page until its script has loaded. On mobile that's most of your LCP. Fix that first. The rest of the list matters a lot less until then.
Want me to walk your developers through it?