Workload before brand
We begin with application type, traffic, operational skill, support needs, and risk tolerance before discussing a provider.
Testing & editorial policy
Evidence first. Limits visible.
Useful performance guidance explains where data came from, what the test conditions were, and where interpretation begins.
Operating principles
How we earn trust.
These rules apply to our tools, educational content, future provider comparisons, and commercial relationships.
Evidence over marketing
Observed data, third-party tool output, provider claims, and editorial judgment are identified as different kinds of evidence.
Lab and field data stay distinct
A controlled run helps diagnose. Aggregated field data describes eligible real-user experience. One does not replace the other.
No invented benchmark results
We do not publish simulated provider rankings as if they were observed tests. Missing evidence is shown as missing—not filled with an estimate.
Trade-offs remain visible
Every category and recommendation includes constraints, questions to verify, and an alternative path.
Commercial relationships stay separate
Affiliate compensation does not determine a Fit Finder result, technical observation, correction, or the outcome of a test.
Current speed-test method
What happens when you press “Run test.”
The current tool is an on-demand interface to the public Google PageSpeed Insights API. SpeedHost Labs does not pretend the result is a proprietary benchmark.
Validate
The browser accepts only HTTP or HTTPS URLs and rejects common private or local-network addresses.
Request
The public URL and selected mobile or desktop strategy are sent to Google PageSpeed Insights only after you submit the form.
Organize
Returned Lighthouse category scores, lab metrics, and audit opportunities are placed into a simpler readable layout.
Interpret
Guidance distinguishes page-level work from possible hosting or backend clues. It does not diagnose a provider from one result.
Important: a lab result is an estimate captured under
specific simulated conditions. It can vary between runs and does not
guarantee search ranking, revenue, conversion, uptime, or the
experience of every visitor.
Provider benchmarks
We publish a ranking only when the dataset can support it.
A future hosting comparison must be repeatable enough to reveal a trend, not just a lucky run.
Multiple runs
Tests need repeated samples over time, with failures and outliers documented instead of quietly removed.
Equivalent workloads
Plans, software, page assets, cache state, traffic pattern, and region must be close enough for a comparison to mean something.
Visible dates
Hosting infrastructure changes. Results need test dates, plan details, locations, and a review schedule.
Availability and errors
Median speed is incomplete without error rate, failed tests, uptime, recovery, and meaningful support evidence.
Commercial context
Free accounts, review access, affiliate links, sponsorship, and provider responses must be disclosed near the evaluation.
No universal winner
A provider can be fast in one region or workload and unsuitable in another. Conclusions stay scoped to the evidence.
Editorial independence
Compensation does not buy the result.
Some pages may eventually include clearly labeled affiliate links. That relationship does not alter tool output or category recommendations.
Read the disclosureCoverage is not for sale
Payment does not guarantee inclusion, a positive review, a score, placement, or removal of legitimate criticism.
Affiliate links are labeled
A nearby or page-level notice will identify a commercial relationship where a qualifying purchase may earn a commission.
Provider claims are attributed
Pricing, feature, uptime, or performance claims supplied by a provider remain provider claims until independently verified.
Corrections welcome
Email a correction
Found an error or missing context?
Send the page URL, the statement in question, and supporting evidence. We review substantive corrections and update material errors.