Halloween is the rehearsal. Black Friday is the horror show.Black Friday season is here. Every slow second costs you sales. Try Nightmare Mode →

Aerial view of the Vitacom building

Vitacom Electronics

Romanian electronics retailer, in business since 1991. Already on Magento and Hyvä, still slow, fragile, and exposed. Then we rebuilt it into a 100 PageSpeed store on eight times less hardware.

100 Performance
97 Accessibility
100 Best Practices
100 SEO
Magento 2
Hyvä Theme
Hyvä Checkout
Liquid Framework

Vitacom live stores

Explore the public storefront and the B2B and export portals we rebuilt.

The rescue

The store we inherited

Vitacom had invested in the right platform. Magento 2 with Hyvä is an enterprise-grade, high-performance choice, and the intent behind it was clear. But over the years, extensions, customizations, and server changes had piled up from different teams with no unified strategy. The result was not a bad platform. It was a good platform buried under years of technical debt.

When we picked it up, the store was on Magento 2.4.6-p1, left unpatched for roughly two years, with a dozen known security holes open. Frontend PageSpeed sat at 12 and server response peaked near 16 seconds. On a Hyvä store. That is not what Hyvä does when it is built correctly.

The store we inherited
The stack promised blazing speed. The build delivered a 12 PageSpeed.

Diagnosis

A single page could trigger 15 million database operations

The audit told the real story. One module that was not even in use fired 9.9 million database queries on every single product view. A product-label module added 5.4 million more. Sorting logic threw 2.2 million. Reindexing had drifted so far that Catalog Search took over two hours, where healthy values are measured in seconds.

This is what the wrong build looks like under the hood. Nothing here was Hyvä. Nothing here was Magento. Every one of these numbers came from how the store had been assembled and configured over time.

A single page could trigger 15 million database operations
Query load on a single product view, before cleanup.
Performance Comparison

Core Web Vitals & Performance Metrics

Google's Core Web Vitals metrics are essential to measuring the quality of user experiences. See how our implementation transformed the site's performance scores.

Before

Time to First Byte (TTFB)

Extremely slow server response

15,814ms
First Contentful Paint (FCP)
Core

Poor - Delayed first paint

2,508ms
Largest Contentful Paint (LCP)
Core

Poor - Delayed loading of main content

2,900ms
Time to Interactive (TTI)

Long wait before usability

11,191ms
Total Blocking Time (TBT)

Excessive main thread blocking

3,045ms
Speed Index (SI)

Slow overall experience

27,539ms
Cumulative Layout Shift (CLS)
Core

Poor - Unstable, shifting interface

0
Overall PageSpeed Score: 12/100

After

Time to First Byte (TTFB)

Near-instant server response

12ms
First Contentful Paint (FCP)
Core

Excellent - Instant first paint

459ms
Largest Contentful Paint (LCP)
Core

Excellent - Fast content rendering

539ms
Time to Interactive (TTI)

Interface usable immediately

623ms
Total Blocking Time (TBT)

Zero main thread blocking

6ms
Speed Index (SI)

Extremely fast visual progression

1,049ms
Cumulative Layout Shift (CLS)
Core

Excellent - Perfect layout stability

0
Overall PageSpeed Score: 100/100

* Core Web Vitals are Google's user-centric metrics that measure real user experience

Good
Needs Improvement
Poor

The method

Secure, simplify, then accelerate

We never bolt performance onto a broken base. Our approach runs in three phases, and the order is deliberate. First we make the platform secure and stable. Then we strip out the bloat and rebuild a clean Magento skeleton. Only then do we introduce the Liquid Framework to deliver the speed. A performance layer on top of a compromised, overloaded setup would only add more risk.

Secure, simplify, then accelerate
Secure and Harden, then Clean and Simplify, then Accelerate.

Phase 1

Secure and harden

We rebuilt production on a purpose-built Magento server, a clean Varnish to Nginx to php-fpm stack, and upgraded Magento from 2.4.6-p1 to 2.4.6-p13 with every missing security patch applied. We closed the exposed test sites, retired TLS 1.0 and 1.1, shut the open FTP port, enforced two-factor authentication, and cleaned up stale admin access.

This step is future-proof by design. Even if Vitacom had chosen to walk away after it, they would have inherited a secure, correctly configured environment instead of a liability.

Secure and harden
Phase 1: from exposed to hardened.

Phase 1

Deploys stopped being a nightmare

Under the old setup, every release was a gamble. Manual deploys broke things, and the client's IT department spent its weeks firefighting instead of building. We put a proper automated pipeline in place: commit, build, test, deploy, with calm and predictable releases. That alone gave the team back their time.

Deploys stopped being a nightmare
Manual chaos replaced with an automated CI/CD pipeline.

Phase 2

Clean and simplify

We removed roughly 180 modules, including about 135 from the single dominant vendor, and rebuilt the custom modules that could silently corrupt or lose data. Search moved to OpenSearch. Heavy, overlapping extension suites were replaced with lean, targeted solutions. What was left is a clean Magento skeleton, ready to scale.

Less is more. Fewer modules mean fewer conflicts, fewer outages, cheaper upgrades, and up to 31,500 euro a year saved on extension licenses alone.

Clean and simplify
240+ modules and years of debt, reduced to a clean skeleton.

The results

Same catalog, same Magento, same Hyvä. A completely different store, on a fraction of the hardware.

New website launch: Q2 2026

99.92%

Server response

Faster TTFB, 16s to 12ms

8x less

Hardware

64 to 8 cores, 256 to 32 GB

12 → 100

PageSpeed

Frontend score

~180

Modules removed

Extension bloat stripped out

15M removed

Database load

All from third-party modules

€31,500/yr

License savings

Extensions retired

Magento 2
Hyvä Theme
Hyvä Checkout
Liquid Framework

Phase 3

Accelerate, on eight times less hardware

On the clean foundation, we rolled out the Liquid Framework and Hyvä Checkout. Then we did something that tells the whole story: we moved the store off its oversized 64-core, 256 GB server and onto a right-sized 8-core, 32 GB machine tuned for Magento. Eight times less hardware, and the server load actually dropped.

Accelerate, on eight times less hardware
Eight times less hardware, around 1,300 times faster response.
Server CPU load dropping after cleanup and the ElasticSuite launch
Server CPU load falling after the cleanup and ElasticSuite launch.
Database CPU load dropping after Liquid Framework go-live
Database CPU load falling after the cleanup and ElasticSuite launch.

From 16 seconds to 12 milliseconds

The headline number is server response time. It went from 15,814 milliseconds to 12. That is a 99.92 percent improvement, roughly 1,300 times faster, on a server with a fraction of the resources.

From 16 seconds to 12 milliseconds
Server response time, before and after.

The transformation (before and after)

The mess we inherited on the left, the store we delivered on the right. Drag to compare.

After

Hyvä Metrics

First Contentful Paint (FCP)

459ms

Largest Contentful Paint (LCP)

539ms

Cumulative Layout Shift (CLS)

0

After: 12 ms TTFB load time
Before

Old Metrics

First Contentful Paint (FCP)

2,508ms

Largest Contentful Paint (LCP)

2,900ms

Cumulative Layout Shift (CLS)

0

Before: 16 s TTFB load time
Hover to see before & after

See the difference

The rebuild turns years of accumulated technical debt into a fast, stable, high-converting store.

Vitacom new homepage
Vitacom new category page
Vitacom new product page

The results

Same stack, a completely different store

Same catalog. Same Magento. Same Hyvä. PageSpeed went from 12 to 100. First Contentful Paint landed at 459ms, Largest Contentful Paint at 539ms, and Time to First Byte collapsed to 12ms. Cumulative Layout Shift hit zero. The platform that could not clear a 12 now runs a perfect PageSpeed, and it does it on a fraction of the hardware.

Same stack, a completely different store
Vitacom, representative page load, before and after.
Real Vitacom PageSpeed report showing 100/100 with green Core Web Vitals
The live result: a real 100 PageSpeed with green Core Web Vitals.

The rescue, in short

Vitacom had the right platform and the wrong build. We secured it, stripped years of bloat, and only then unleashed the speed.

Inherited a Magento 2.4.6-p1 store on Hyvä, unpatched for two years, sitting at a 12 PageSpeed with server response peaking near 16 seconds.

Found 240+ third-party modules, 51 custom modules, 12 open security holes, and single pages triggering up to 15 million database operations.

Rebuilt in three phases: secure and harden, clean and simplify, then accelerate with the Liquid Framework and Hyvä Checkout.

Delivered a 100 PageSpeed and a 12 millisecond server response, on a server with eight times less hardware.

Our approach (secure, simplify, accelerate)

1

Secure and harden

Rebuilt production on a purpose-built Magento server, upgraded 2.4.6-p1 to 2.4.6-p13 with every missing patch, closed the exposed test sites, retired TLS 1.0 and 1.1, and enforced proper access controls.

2

Clean and simplify

Removed roughly 180 modules, rebuilt the custom modules that could corrupt data, moved search to OpenSearch, and cut the query storms down to a clean Magento skeleton.

3

Accelerate

Rolled out the Liquid Framework and Hyvä Checkout on the clean foundation, then right-sized the server from 64 cores to 8. Eight times less hardware, roughly 1,300 times faster.

Why it matters

The technology was never the problem

This is the case study every merchant should read before choosing a Magento partner. Hyvä and Magento deliver exactly what they promise. Whether you get a 12 or a 100 comes down to who builds it. We are not an agency that uses Magento now and then and invoices for it. We are Hyvä and Magento community members, and we contribute to the core code that makes the ecosystem better. That is the difference between a store that fails on great technology and one that flies on it.

The technology was never the problem
Eight times lower server cost, licenses removed, 100 PageSpeed on the same stack.

Is your Magento store slower than it should be?

If you are on Magento or Hyvä and not seeing the speed you were promised, the technology is probably fine. The build might not be. Let us take a look.

Not ready to talk yet?

Get one email a month with results like these and the fixes behind them.

One email a month. Unsubscribe in one click. Privacy policy

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.