Disclosure: We may earn a commission if you purchase hosting through links on this page, at no additional cost to you. All speed data is independently gathered.

How We Test Web Hosting in Australia

This page explains exactly what we test, how we test it, and what we deliberately leave out. It also explains what we have not finished testing yet, because we would rather publish real data late than fake data on time.

Current status (June 2026): our controlled Australian speed benchmark is in progress. Where a review says "benchmark in progress", it means exactly that: no measured figure is published until our own test produces it. Any performance commentary in the meantime is structural (server location, stack, physics) and labelled as such.

Speed Testing - Time to First Byte (TTFB)

TTFB is the time between a browser sending a request and receiving the first byte of a response. It is the clearest single measure of server performance - it strips out page weight, image sizes, and front-end code, and tells you how fast the server itself responds.

Our test setup

Identical test page on every host. We deploy the same lightweight HTML test page to each hosting account we own, with no CDN and all host-level caching disabled. We test the bare origin, because a CDN edge cache can make any host look fast while hiding slow server responses.
Real Australian probes in four cities. Measurements run through the Globalping probe network from Sydney, Melbourne, Brisbane and Perth - real machines on real Australian consumer and business connections, not a single data-centre VM.
Repeated runs, honest statistics. Each benchmark cycle takes repeated measurements per city and we publish the median and the 95th percentile, not a cherry-picked best run.
Hosts we pay for. Benchmarked hosts are accounts we purchased ourselves. Where we have not bought an account, we say so and do not publish invented numbers for it.

Uptime Monitoring

Test sites are monitored with third-party uptime checks on a rolling window, checked every 5 minutes from Australian nodes. Results are published per host once a full monitoring window completes. A host that claims 99.9% uptime but routes that claim through their own dashboard does not impress us; until our own monitoring window completes, the only uptime figures you will see in our reviews are the provider's published SLA, labelled as such.

Support Testing

We contact support via live chat during AEST business hours with a standard technical question, and time from first message to a useful response - not just "Hi, how can I help?" but an actual answer. Response-time figures appear in a review only after we have run this test for that provider; otherwise we describe support structure (channels, hours, where the team is based) without inventing times.

Pricing Verification

On every review we record the advertised first-term price, the renewal price, and the minimum contract length required, because the renewal price is the real price. Prices change frequently and some providers bill in USD; where a figure is approximate or converted we mark it, and you should always confirm the current price on the provider's own checkout before buying.

Re-test Cycle

Once the benchmark is live, every host goes onto a recurring re-test cycle. Infrastructure changes, and we do not freeze results from a single good day. Each results table will display the date of its most recent test run.

How Our Review Scores Work

Every review carries a score card with component scores out of 10 and an overall score. The rules:

Component scores are editorial judgments against defined criteria - infrastructure quality and location for Speed, redundancy and SLA for Uptime, channels, hours and team location for Support, renewal-price honesty for Value.
No component scores a 10 until our own measurement backs it. Speed and Uptime scores are provisional until the benchmark and monitoring windows complete for that host.
The overall score is the mean of the components, rounded to the nearest 0.5. We allow ourselves at most a 0.5-point editorial adjustment either way, and when we use it, the review says so and explains why.
Scores change when facts change. A pricing rise, an infrastructure change or a benchmark result can move a score; the review's last-updated date reflects the most recent assessment.

What We Deliberately Leave Out

No scores for things we have not measured. If a number is not ours, we either cite its source (a provider SLA, a published spec) or we leave it out.
No "tested" badges on desk research. Comparing plans, pricing and policies is research, not testing, and we label it accordingly.
No pay-to-play rankings. Commission rates do not move a host up or down our rankings. Several hosts we recommend pay us nothing.

Affiliate Disclosure

Pick a Host earns commissions when you sign up with some providers through our links. That funds the test accounts. It does not change the methodology above, and it never overrides a measured result. Full details on our affiliate disclosure page.