Performance audit
What slows your application down, in black and white.
Your website or web application loads too slowly, but nobody knows exactly why. In the performance audit I measure it under realistic conditions and show which causes account for the largest share and which measures pay off first.
What the performance audit examines
Load time and Core Web Vitals
I measure LCP, CLS and INP on your most important pages, on an average smartphone and a throttled connection, not just on a developer machine.
Delivered data
I check what the browser has to load: JavaScript, CSS, images, fonts and third-party scripts, and how much of it is not needed for the first impression at all.
Server and delivery
I check response times, caching, compression and the order in which resources are loaded. With load tests in k6, I also measure how many users your infrastructure can serve at the same time and when it starts to slow down.
What you have in hand after the performance audit
Measurements and load limit
The metrics of your most important pages as a baseline that every later improvement can be measured against, and how many users your application handles at the same time before it slows down.
Prioritized findings
Every cause with evidence, impact and estimated effort, sorted by what pays off most.
Action plan
Concrete steps in a sensible order that your team can implement itself or that I implement for you.
Closing call
We go through the results together and answer your team's questions.
Example finding from a performance audit
This is how every finding in the report is structured. The example is made up and shows how specific the results are.
Finding
The hero image of the home page is delivered as a 2.4 MB PNG and only loads after a script requests it.
Impact
On a smartphone the largest visible content appears after 4.8 seconds. At most 2.5 seconds is considered good.
Recommendation
Deliver the image in a modern format and in fitting sizes and reference it directly in the HTML. Effort about an hour, expected effect around two seconds.
How the audit works
Request the audit
You pick your audit and briefly describe your application.
Fixed price and analysis
You get the fixed price for your application, then I review it.
Findings and action plan
You receive prioritized findings and a plan for the next steps.
Why a performance audit despite AI tools?
Tools like Lighthouse and AI assistants produce long lists of hints within seconds. I use them myself to evaluate measurements and search code faster. Which three out of fifty items matter for your users and your revenue, what can be changed in your architecture with reasonable effort and what you can safely ignore, none of these lists will tell you. That assessment is the actual result of the audit.
Frequently asked questions about the performance audit
What do you need from me for the performance audit?
The address of the application and, if it sits behind a login, a test account. For the load tests with k6, I need a test environment that resembles your production as closely as possible, or a time window agreed with you. Access to the source code and existing measurement data helps but is not required.
What does the performance audit cost?
The audit comes at a fixed price from €1,500. I tell you the exact price before I start, it depends on the size of your application. If a follow-up project comes out of it, the audit price is credited towards it.
Do you also implement the measures?
Yes, if you want me to. The action plan is written so that your team can implement it itself. If I take over the implementation, the audit price is credited towards the project.
Will the application stay fast after the audit?
Only if someone keeps an eye on it. That is why the action plan recommends limits that are checked automatically with every change. The case study on the performance budget of this website shows what that looks like.
Request the audit or talk first
Request your audit directly and you get the fixed price and time frame, with no obligation. Not sure which audit fits? Then we sort that out in a free initial call.
Problem
Your application keeps users waiting
Long loading times, janky interfaces and poor Core Web Vitals. I find the causes and make your application measurably faster.
Case study
A website that has to stay fast
Three limits for loading time, layout and page weight help decide here whether a version goes live. This is how the performance budget of this website is built.