WordPress Hosting Speed Test: 10 Hosts Benchmarked (2026)

Why hosting benchmarks are usually misleading

Most WordPress hosting comparison articles have a conflict of interest. Affiliate commissions for managed WordPress hosting average €50-200 per signup. This creates pressure to rank higher-commission hosts above lower-commission ones regardless of actual performance.

The tests in this article were run independently. No host in this comparison has an affiliate relationship with this article. Benchmark data comes from real test sites and real client sites, not sponsored environments.

Test methodology

Test environment:

• WordPress 6.7 installed fresh on each host
• Twenty Twenty-Four theme (block theme, no plugins except Hello Dolly which was immediately deleted)
• No caching plugins active for uncached tests
• WP Super Cache activated with default settings for cached tests
• PHP 8.3 on all hosts where available (8.2 otherwise)

Measurement method:

• TTFB measured using browser DevTools Network panel from Sofia, Bulgaria
• 10 measurements per host taken over 3 different days and times
• Median value reported (not average, to reduce outlier impact)
• Lighthouse Performance score: PageSpeed Insights API, simulated mobile, Slow 4G

Why Sofia, Bulgaria: The target audience for this site is predominantly EU-based. Sofia is a mid-sized EU city without the benefit of being in a major CDN hub. This gives a realistic view of performance for typical Eastern/Central European visitors.

Results: TTFB and Lighthouse scores

Uncached PHP requests

This is the baseline - PHP executing, database querying, WordPress bootstrapping without any caching assistance.

With full-page caching active

These numbers reflect what most real WordPress sites serve to most visitors: cached HTML responses.

Lighthouse Performance Score (mobile, cached)

Host-by-host analysis

Hetzner CX32 with CloudPanel

The best price-to-performance option in this test. A Hetzner CX32 server (4 vCPUs, 8GB RAM, NVMe SSD, Frankfurt) with CloudPanel installed costs €6.57/month and runs 10–15 WordPress sites.

CloudPanel installs a full stack — Nginx, PHP-FPM (multiple versions), MySQL or MariaDB, Varnish Cache, and Redis — all included at no extra cost. New WordPress sites are provisioned in under a minute via the UI. Varnish Cache delivers 35ms cached TTFB; Redis object cache is available for WooCommerce session management without the extra €14/month Cloudways charges for it.

This is not managed hosting. You are responsible for OS security updates, PHP upgrades, and monitoring. Budget 1–2 hours per month.

Verdict: Best overall choice for technically confident agencies managing multiple sites. The per-site cost works out under €1/month. See the full Hetzner + CloudPanel setup guide.

Cloudways Vultr High Frequency

Second-fastest in the managed hosting category. The Vultr HF plan uses high-clock-speed Intel processors optimised for single-threaded workloads - exactly what PHP execution is.

The Cloudways management layer adds convenience without adding meaningful overhead. Deploying a new WordPress site takes 3 minutes. Cloning a site for staging takes 5 minutes. These developer-quality tools are why agencies choose Cloudways.

Varnish caching (32ms TTFB) is slower than Kinsta's Nginx cache (28ms) in direct comparison, but the difference is imperceptible in real-world usage. The Breeze plugin handles WooCommerce exclusions correctly without manual configuration.

Cost advantage: Multiple sites on one server. A Vultr HF 2GB server (€25/mo) handles 5-8 low-traffic WordPress sites. This makes the per-site cost far lower than single-site managed hosting.

Verdict: Best managed hosting for agencies managing multiple sites on a budget.

Rocket.net

Rocket.net is worth noting separately because its architecture is different from every other host here. Every site is deployed behind Cloudflare Enterprise, with full-page caching at Cloudflare's edge nodes. The 14ms cached TTFB is better than Kinsta's 28ms because it is served from a Cloudflare edge node close to the visitor, not from a single origin server.

Uncached TTFB (214ms) is comparable to Kinsta. The edge caching advantage only applies to cache hits - which is most traffic on a typical content site.

Drawback: €30/mo is expensive for a single site compared to Cloudways. The Cloudflare Enterprise edge caching is the differentiation.

Verdict: Strong option if you want Cloudflare Enterprise without configuring it yourself.

Kinsta

Kinsta's uncached TTFB (218ms) and cached TTFB (28ms) are strong and consistent across all 10 measurements. The maximum TTFB of 248ms on an uncached request shows tight performance variance - unlike shared hosts where the max can be 2-3x the median.

The My Kinsta dashboard is the best management interface tested. PHP error logs, cache hit rates, and resource usage are all visible per-site. For production site management, this observability has value beyond the raw performance numbers.

Drawback: Per-site pricing. A Kinsta Starter plan (25,000 visits/month) costs €35/mo for one site. Cloudways at €14/mo hosts multiple sites on one server.

Verdict: Premium option justified for client sites where infrastructure quality and support matter most.

WP Engine

WP Engine's TTFB (261ms uncached, 41ms cached) is slower than Kinsta and Cloudways but still respectable. The higher uncached TTFB correlates with their infrastructure running on older-generation compute.

EverCache, WP Engine's caching system, handles WooCommerce cart exclusions and is deeply integrated with WordPress. But at 41ms cached TTFB vs Kinsta's 28ms, the performance advantage Kinsta has is tangible on high-traffic sites.

Where WP Engine wins: Developer workflow features. Genesis framework, Smart Plugin Manager, and Atlas headless support are differentiators for agencies with enterprise clients.

Verdict: Right choice for enterprise clients who need the workflow features. Not the best choice for performance-sensitive WooCommerce stores on a budget.

SiteGround Go Geek

SiteGround's 382ms uncached TTFB is the first result in this test that would noticeably affect a WooCommerce store's conversion rate. The 87ms cached TTFB is slower than competitors but functional for content sites.

The Sofia datacenter is SiteGround's home datacenter. For Bulgarian clients, this geographic proximity should be an advantage. The benchmark was run from Sofia to SiteGround Sofia, giving SiteGround the best possible conditions. The 382ms result is therefore not masked by network latency.

Verdict: Acceptable for brochure sites with audiences close to EU datacenters. Not recommended for WooCommerce.

Hostinger Business

Hostinger's 428ms uncached TTFB and 148ms cached TTFB reflect shared hosting infrastructure. The LiteSpeed server helps (cached TTFB would be much worse without it) but shared CPU and I/O contention limit performance.

At €3.99-8.99/mo, Hostinger is the cheapest credible option in this test. For a personal blog or static business site where slow page loads have no conversion impact, Hostinger is fine.

Verdict: Suitable for very low-traffic sites where cost is the primary concern.

Bluehost Choice Plus

Included as a comparison point because Bluehost appears prominently in many hosting comparison articles (driven by high affiliate commissions). The 891ms median TTFB with US servers is not competitive for EU audiences. Even with a CDN for static assets, dynamic WordPress pages would load slowly for European visitors.

Verdict: Not recommended for EU-targeting WordPress sites.

What the numbers mean in practice

A 200ms vs 400ms TTFB difference sounds small. The actual impact:

On a blog: Visitors who are already on your page do not notice a 200ms difference in TTFB. It affects time-to-first-byte, which affects the start of rendering, which affects LCP by roughly the same amount. A 200ms worse TTFB often means a 200-300ms worse LCP.

On a WooCommerce store: Cart and checkout pages cannot be fully cached. Every checkout step is a dynamic request. On a checkout flow with 3 steps, 200ms worse TTFB = 600ms added to checkout time. This is measurable in cart abandonment rates.

On a high-traffic site: At 10,000 concurrent users, the difference between 200ms and 400ms TTFB means your server is handling twice as many in-flight requests. This accelerates the point at which you hit resource limits.

The recommendation

Best overall price/quality — Hetzner CX32 + CloudPanel: 35ms cached TTFB, 155ms uncached, €6.57/month for a server that runs 10+ sites. Nginx + Varnish caching, Redis included free, one-click WordPress installs. Nothing in this test matches the price/quality ratio. Full setup guide here.

Best managed hosting — Rocket.net: 14ms cached TTFB via Cloudflare Enterprise edge. The fastest managed host in this test. €30/mo per site.

Best value managed for agencies — Cloudways Vultr HF: Per-server pricing, strong performance (32ms cached, 198ms uncached), multi-site economics. The right choice for agencies managing 5+ client sites on managed infrastructure.

The benchmarks under real-world conditions

The clean test environment numbers above represent best-case performance. Real WordPress sites have plugins, images, and dynamic content. Here is how each host performs under more realistic conditions.

Test 2: WooCommerce store with 200 products

Same hosts, but with WooCommerce 9.x installed, Storefront theme, 200 products imported, and WP Rocket caching configured correctly with cart exclusions.

The uncached checkout TTFB numbers are what matter for conversion rates. Checkout pages cannot be cached - they are dynamic, session-specific. The 195ms difference between Cloudways (215ms) and SiteGround (410ms) on checkout is meaningful and shows up in cart abandonment data.

Test 3: Under simulated traffic load

Using k6 to send 50 concurrent requests per second for 30 seconds, simulating a promotional email blast or press coverage traffic spike:

SiteGround and Hostinger show significant performance degradation under load. The error rates (504 Gateway Timeout responses) are a known characteristic of shared hosting under concurrent traffic.

Managed VPS and container-based hosting (Kinsta, Cloudways) handles traffic spikes without error rate increases because resources are dedicated rather than shared.

PHP version impact on performance

All tests above used PHP 8.3. PHP version has a significant and often-overlooked impact on WordPress TTFB:

Upgrading PHP version alone can improve TTFB by 20-30ms without any other changes. Test on staging first since older plugins may not be compatible with PHP 8.x.

Check your current PHP version: WordPress Admin > Tools > Site Health > Info > Server.

Datacenter location and EU audiences

All measurements in this test were taken from Sofia, Bulgaria to Frankfurt datacenters. For different EU audiences, the proximity of the datacenter affects the base network latency:

CDN services (Cloudflare, BunnyCDN) eliminate geographic penalties for static assets - images, CSS, and JavaScript are served from the nearest edge node. Only uncached dynamic PHP requests need to reach the origin server.

For sites targeting multiple EU regions, Frankfurt is the best single-datacenter compromise, combined with a CDN for static assets.

Server configuration that affects scores

Beyond the host, server configuration choices affect TTFB significantly. Two settings with the largest impact:

OPcache configuration:

OPcache caches compiled PHP bytecode in memory. Without it, PHP compiles WordPress's ~400KB of PHP files on every request. Most managed hosts enable OPcache correctly. On self-managed VPS, verify OPcache is active and configured with adequate memory:

; php.ini
opcache.enable=1
opcache.memory_consumption=256
opcache.interned_strings_buffer=16
opcache.max_accelerated_files=10000
opcache.revalidate_freq=60

With OPcache properly configured, PHP execution time for a typical WordPress page drops from 180ms to 60-80ms.

Nginx worker configuration:

On Nginx servers, worker processes and connection limits affect performance under load. The default configuration on most VPS setups is conservative:

# nginx.conf
worker_processes auto;  # one worker per CPU core
worker_connections 1024;
keepalive_timeout 65;

For a 2-core VPS running multiple WordPress sites: worker_processes 2 and worker_connections 2048 is a reliable starting point. Managed hosts (Cloudways, Kinsta) handle this automatically.

What these benchmarks mean for ranking

Google uses Core Web Vitals as a ranking signal in its Page Experience algorithm. The specific signals: LCP, CLS, and INP measured from real user data (CrUX - Chrome User Experience Report).

The connection to hosting speed:

• Faster TTFB → faster LCP (LCP cannot be faster than TTFB)
• LCP under 2.5s → "Good" LCP in CrUX
• "Good" LCP contributes to the Page Experience ranking signal

The competitive advantage from hosting speed in search rankings is real but modest. Google's documentation describes Page Experience as a tiebreaker when content quality is similar between pages. A site on Kinsta does not outrank a more authoritative site on Hostinger.

The more direct benefit of fast hosting is conversion rate and user experience - measurable in bounce rate, session duration, and checkout completion for e-commerce.

Conclusion: the right host for your situation

After running these tests across multiple rounds, the recommendations are clear:

For EU agencies managing client sites: Cloudways Vultr HF for cost efficiency, Kinsta for premium clients where support quality justifies the cost.

For individual site owners: Hetzner VPS with CloudPanel if you are technically comfortable managing a server — full setup guide here. SiteGround Go Geek if you want managed hosting at a reasonable price point.

For WooCommerce stores: The uncached TTFB numbers matter most. Any host above 400ms uncached TTFB will affect checkout conversion rates. Cloudways and Kinsta are the clear choices for stores where conversion rate has a direct revenue impact.

For enterprise or high-traffic sites: Kinsta or WP Engine, where SLA guarantees, auto-scaling, and premium support justify the higher cost.

Related reading

• Best WordPress Hosting 2026
• Cloudways vs Kinsta: Which Managed Host Is Right for Agencies?
• WordPress Core Web Vitals Fix Checklist 2026