Core Web Vitals
Improve LCP, CLS and INP — the speed metrics Google ranks on.
// Growth · Performance
A faster shop sells more
Speed up a slow Shopware shop — better Core Web Vitals, caching, database and template tuning for higher rankings and conversion.
// What you get
How we make shops fast
Caching & HTTP
HTTP cache, Redis and Varnish tuning for instant repeat loads.
Database tuning
Find and fix slow queries and indexes that drag the shop down.
Asset optimisation
Compress and lazy-load images, CSS and JS for lighter pages.
Template review
Remove render-blocking and heavy template logic that slows pages.
Hosting & CDN
Advise on hosting and set up a CDN for fast global delivery.
// Why it matters
Speed is money and ranking
Every extra second of load time costs conversions, and Google ranks slow shops lower. Codewerk Solutions runs a full Shopware performance audit — Core Web Vitals, caching, database queries, template logic, images and hosting — then fixes the bottlenecks that actually matter. We measure before and after, so you can see the improvement in real numbers.
The result is a shop that loads fast, ranks better and converts more visitors into buyers — one of the highest-return investments you can make in an existing store.
// In detail
Technology, performance and SEO
What we actually do under the hood — so you can judge the work, not just the promise.
Technology & architecture
We start by measuring, not guessing: TTFB, the query profiler on a production-sized catalogue, and the real field data from the Chrome UX Report. Then we fix causes in order — HTTP cache, N+1 queries in plugins, missing indexes, synchronous ERP calls during render, and shared hosting CPU.
Caching & infrastructure
HTTP cache and reverse proxy in front, Redis for sessions and object cache behind, Elasticsearch when the catalogue justifies it. Cache invalidation is tagged, so a price change invalidates exactly what it should — and you never have to flush everything at 10 a.m. on a Monday.
SEO & discoverability
Core Web Vitals are the overlap between performance and SEO. We target LCP under 2.5s, CLS under 0.1 and INP under 200ms on the pages that actually carry revenue — the category listing, the product page and the cart, not the homepage everyone benchmarks.
// FAQ
Frequently asked questions
On an unoptimised Shopware shop, a 2–4× improvement in TTFB is common and largely comes from caching and fixing a handful of bad queries. We tell you the expected range after a paid audit, not before.
Below roughly 10,000 simple products, usually not — the database is fine and Elasticsearch is a second system to operate for nothing. Above that, or with heavy faceting and typo tolerance, it pays for itself quickly.
Caching is where it can, which is why prices and stock are never cached for minutes. Everything is rehearsed on staging with real data, and we agree in writing what must always be live rather than cached.
Because hosting is rarely the bottleneck. In most shops we audit, the time goes into plugins that query the database once per product in a loop, missing indexes, a synchronous ERP call during page render, and an HTTP cache that is effectively switched off. A bigger server makes that cheaper to survive, not faster. Changing host before measuring usually just relocates the problem.
Less than people think. PageSpeed Insights runs a lab test on a throttled phone; Google ranks on field data from real visitors, which can look quite different. Chase the field metrics — LCP, CLS and INP on your category and product pages — and the score follows. Optimising for the score itself leads to tricks that help the test and nobody else.
For images and static assets it is cheap and worth having. As a cure for a slow shop it mostly misses the point: if your customers are all in the DACH region and your server stands in Germany, distance was never the problem. A CDN cannot make a slow database query fast. Fix the origin first, then put a CDN in front of it.
The audit is a matter of days and gives you a list ranked by expected effect. The fixes vary: caching and index work often land within a week or two and produce most of the gain, while replacing a badly built plugin or pulling an ERP call out of the render path takes considerably longer. We sequence it so the cheap wins ship first, not last.
Let's talk about your project
Send a short brief — we reply within one business day.