Performance

Slow has a cause. Find it, then fix it.

Most speed work is guesswork with a caching plugin attached. We measure the site the way Google does, on a throttled mobile connection, then fix the things the measurement actually blames.

Core Web Vitals
Real mobile testing
Caching and assets
Diagnosis first
Built differently

A score is not the problem. Whatever is causing it is.

Most performance work starts at the wrong end. A caching plugin goes on, the desktop score jumps, and the site still feels slow on a phone on mobile data, which is where most of your visitors actually are.

We start with measurement, on a throttled mobile run, then read what it blames. Usually it is a handful of specific things: the wrong image being preloaded, a font library loaded for four icons, a theme shipping CSS for pages that do not exist, a database nobody has cleaned in years.

Occasionally the cause is not in the site at all. We have found platforms sitting in front of a site at the network level, invisible in the plugin list, adding seconds before WordPress even runs. You do not find that by installing another optimization plugin.

Plugin and hope

A caching plugin installed, the desktop number improves, and the phone experience does not change at all.

Score chasing

Points collected against a checklist while the thing your visitors actually wait for stays exactly as slow.

Diagnosed

Measured on a throttled mobile run, causes identified one by one, and each fix verified against the number it was supposed to move.

What's included

What a performance job covers.

The output is not a score. It is a list of causes, the fixes that address them, and a second measurement that shows which ones worked.

A measured starting point

A throttled mobile Lighthouse run and field data, recorded before anything is touched, so improvement is provable rather than claimed.

Core Web Vitals, not just a score

LCP, CLS and INP treated as separate problems with separate causes, because that is how they behave.

Caching and delivery configured properly

Page, asset and object caching set for the site rather than switched on and left at defaults.

Images, fonts and scripts put on a diet

Oversized images served in modern formats, icon libraries replaced by the handful of icons actually used, and scripts loading on pages that never need them.

A database not carrying ten years of rubbish

Revisions, expired transients, orphaned meta and abandoned plugin tables cleared, then indexes checked on the queries that matter.

The server checked, not assumed

PHP version, memory, object cache and where the site physically sits are all part of the number. If the host is the ceiling, we say so rather than tuning around it.

Measured with

Lighthouse

Field data

Core Web Vitals

Caching

WP Rocket

Images

WebP and AVIF

Report

Before and after

Performance

What we actually measure.

Four numbers, four different causes. Treating them as one score is why most speed work fails.

LCP

The thing you wait for

The largest image or block of text on screen. Usually the biggest single win, and usually the wrong element is being preloaded.

INP

Whether it responds

How quickly the page reacts when someone taps. Almost always a JavaScript problem, not a hosting one.

CLS

Whether it stays still

Content jumping around while the page loads. Cheap to fix and disproportionately annoying.

TTFB

How fast the server answers

Hosting, PHP and the database. If this is bad, nothing on the front end will save it.

How we work

How the work runs.

Nothing is changed before there is a number to compare it against.

Measure

A throttled mobile Lighthouse run plus field data, recorded as the baseline.

Diagnose

Each metric is traced to its actual cause rather than to a generic checklist item.

Agree the list

You see what we found, what each fix is worth, and what we recommend leaving alone.

Fix

Changes are made one group at a time, so it stays clear which one moved the number.

Re-measure

The same test, run the same way, so the before and after are comparable.

Hand back

You get the before and after plus the list of what changed, so the next person to touch the site knows what not to undo.

Recent work

Sites we have made faster.

All projects
Šumi Gorenjka d.o.o.
The Sumi homepage, with a red hero reading Delimo sladke trenutke above a row of Sumi sweet packets

Šumi Gorenjka d.o.o.

A heritage Slovenian confectionery brand, built from scratch and translated for two markets.

Read the case study
Why Jaha Web

The number has to survive contact with a real phone.

Measured the way Google measures

A throttled mobile run and field data, not a desktop score that looks good in a screenshot.

Causes, not a checklist

Every item we fix is one the measurement blamed. We do not sell you work the site does not need.

We tell you when the ceiling is the host

Some sites cannot be fixed by tuning. If the hosting is the limit we say so, instead of charging to tune around it.

Frequently asked

Before we start.

What the work involves, what it can achieve, and where the limits are.

What does performance work cost?

The free audit tells you whether there is anything worth doing. If there is, the work is quoted as a fixed price once we know what is causing the problem. A five-page site and a fourteen-hundred-product shop are not the same job, and a published number would be wrong for one of them.

Will you just install a caching plugin?

No. Caching is one item on a list that the measurement produces. If caching were the whole answer, your host would have done it already.

How much faster will it get?

We tell you after the baseline, not before. What is achievable depends on what is causing the problem. Some causes are cheap to fix and some are the hosting itself.

Do I need to change hosting?

Sometimes. If the server is answering slowly there is no front-end change that hides it. We measure that first and tell you plainly rather than charging to tune around it.

Is a score of 100 the goal?

No. The goal is the site feeling fast on a phone on mobile data. A perfect score on a page nobody waits for is worth nothing.

Will it stay fast?

Not by itself. Performance decays as content, plugins and tracking scripts get added over time. What we hand over includes what was changed, so it stays clear what should not be undone.

Do you work on sites you did not build?

Most of this work is on sites built by someone else. We do not need to have written it to measure it.

Can you look before I commit to anything?

Yes. The free audit covers a first measurement and tells you whether there is anything worth doing at all.