Disclosure: We may earn commissions from qualifying purchases on our affiliate links; these fees never influence our rigorous testing, ensuring our SiteGround speed benchma
SiteGround's performance stack is unusually observable: its NGINX Dynamic Cache returns headers that identify a HIT, MISS, or BYPASS. That makes it possible to distinguish fast cached delivery from real WordPress generation instead of reporting one unexplained load-time number.
What is being tested
SiteGround StartUp targets one modest site. GrowBig adds multiple sites, staging, on-demand backups, and broader tooling. GoGeek increases resources and adds developer-oriented features. Introductory prices are usually attractive, while renewals can reach the high teens or more monthly; check current pricing.
All plans use Site Tools rather than cPanel. SiteGround's Speed Optimizer plugin integrates WordPress with static delivery, Dynamic Cache, front-end optimization, and optional Memcached. The provider states that Dynamic Cache is enabled by default on its servers.
Editor's Pick. Our team's current top recommendation for this category. (Affiliate link coming soon — we only link programs we've vetted.)
A controlled WordPress fixture
Use current WordPress, the default block theme, one optimized hero image, several post cards, and no third-party scripts. Keep total transferred assets around 1 MB. Use identical content, PHP version, and plugins for competitor tests.
Record the plan, selected data center, test date, PHP version, Speed Optimizer version, CDN state, and Memcached state. GrowBig and GoGeek have different resources; labeling every result simply “SiteGround” prevents replication.
Confirm cache status first
SiteGround documents these response-header states:
- x-proxy-cache: HIT means the request was served from Dynamic Cache.
- x-proxy-cache: MISS means the page was generated or not currently cached; a repeat may become a hit.
- x-proxy-cache: BYPASS means the request was excluded, commonly because of cookies, login state, or route rules.
Use browser developer tools or a command-line header request. Purge through Site Tools > Speed > Caching or the Speed Optimizer plugin. SiteGround notes that browser-specific cache variation can produce separate entries.
The benchmark matrix
| Run | Cache state | Repetitions | Purpose |
|---|---|---|---|
| Homepage warm | HIT | 20–30 | Reverse-proxy delivery consistency |
| Interior post warm | HIT | 20–30 | Normal anonymous reading |
| Homepage after purge | MISS | 10 purge cycles | PHP/database generation |
| Logged-in page | BYPASS | 10–20 | Dynamic WordPress responsiveness |
| Gradual homepage load | HIT | 1–25 virtual users | Cached throughput and errors |
| Safe dynamic endpoint | BYPASS | 1–10 virtual users | Resource behavior |
Do not test wp-admin or checkout at high concurrency on a live production account. Use staging, synthetic data, and provider permission.
TTFB and percentile reporting
Run requests from a region near the selected data center and from the audience's region. Report median, p75, p95, and maximum. A near-region cache HIT below roughly 200–400 ms is a useful small-site target, but no threshold proves quality by itself.
Connection reuse can make later requests faster; say whether the tool opens new connections. DNS and TLS may or may not be included in “TTFB.” Retain raw timing fields so readers can compare definitions.
MISS results expose application generation. Repeat purge cycles because one request can be affected by background maintenance. BYPASS results are especially important for WooCommerce and membership sites, where logged-in traffic cannot rely on full-page cache.
Browser performance
Run WebPageTest or Lighthouse with a specified mobile device, network profile, and location. Use at least three runs and compare medians. Capture LCP, INP or lab responsiveness proxies, CLS, total blocking time, transfer size, and request count.
Google's good field thresholds are LCP at or below 2.5 seconds, INP at or below 200 ms, and CLS at or below 0.1 at the 75th percentile. These are acceptance thresholds, not SiteGround benchmark claims.
If TTFB is fast but LCP fails, optimize the hero asset and render path. If total blocking time is high, reduce page-builder and marketing JavaScript. Hosting cannot execute browser code faster after it has been downloaded.
Load-test interpretation
Ramp virtual users slowly and add realistic pauses. Track requests per second, median, p95, error percentage, and SiteGround resource statistics. Validate response content so a fast 503 page is not counted as success.
Dynamic Cache should allow a small account to serve many more anonymous page views than uncached requests. A BYPASS curve that degrades earlier does not mean caching is dishonest; it identifies the workload boundary. Report the point where p95 rises sharply or errors begin rather than continuing until suspension.
SiteGround publishes visit estimates as planning guidance, not hard traffic caps. CPU seconds and executions depend on what each visit does. One bot crawling expensive search URLs can consume more than many cached readers.
Memcached and database work
Memcached stores reusable objects from database queries. It can help complex navigation, WooCommerce, and large WordPress databases, but it is not full-page cache. Enable it through Site Tools and Speed Optimizer when appropriate, then compare repeatable BYPASS tests.
If improvement is negligible, inspect query count and plugin behavior rather than assuming more caching is always better. Object-cache churn or poorly chosen keys can provide little benefit. Purging Memcached in Site Tools affects all sites in the account, according to SiteGround's current documentation.
SiteGround CDN testing
Test with the CDN off and on from multiple regions. The CDN should improve distance-sensitive static assets and any eligible cached content. Confirm which hostname and records are active, purge after deployment, and do not compare a CDN-enabled run with an origin-only competitor.
CDN latency can occasionally worsen near-origin tests because it adds a network hop. The relevant result is the audience-weighted geographic performance, not the fastest single lab location.
Competitor context
Hostinger integrates LiteSpeed and often costs less on a long promotional term. Bluehost combines managed page cache, Redis object cache, and Cloudflare with beginner-oriented support. Rocket.net offers a more extensive managed edge stack at a much higher monthly price.
SiteGround's strength is the combination of visible cache status, Speed Optimizer integration, staging on GrowBig, and support. Its main weakness is renewal price. A hobby blog may not use enough of the toolkit to justify the second term.
Pros and cons
Pros
- Cache HIT, MISS, and BYPASS are exposed through a documented header.
- NGINX Dynamic Cache is enabled by default.
- Speed Optimizer coordinates WordPress and platform cache.
- GrowBig adds staging and on-demand backups.
Cons
- Renewal pricing is much higher than the promotion.
- StartUp permits only one site.
- Cached results can hide weaker dynamic performance if methodology is poor.
- Proprietary Site Tools differs from cPanel.
FAQ
Is SiteGround fast for WordPress?
Its NGINX Dynamic Cache can serve anonymous pages quickly, and the WordPress integration is mature. Dynamic performance still depends on plan resources, plugins, database work, and data-center distance.
Should I enable Memcached?
Use it for database-heavy or dynamic WordPress and measure the effect. It does not replace page caching and may provide little visible change on a simple cached blog.
Why does the header say BYPASS?
Login cookies, excluded routes, browser-specific variation, or cache-control instructions can bypass Dynamic Cache. Check SiteGround's current exclusion rules before diagnosing a fault.
Which plan should a business test?
GrowBig is the practical baseline because staging, on-demand backups, and broader site capacity support maintenance. GoGeek may suit higher resource needs; compare current allocations.
Verdict
SiteGround is easy to benchmark honestly because its documented cache header exposes what served each response. The platform is strong for owners who will use Dynamic Cache, Speed Optimizer, staging, and backups, but renewal price must be justified. Test HIT and BYPASS p95 under realistic traffic; keep it when both meet the site's workload, not because one cached screenshot looks fast.