Download speed
Estimates how quickly the browser receives test data. Wi-Fi, device performance, browser overhead, VPNs, congestion and the selected route can all affect the figure.
Transparent testing and research
What LinkSpeed measures—and what the results do not prove
LinkSpeed is a browser-based diagnostic service. A result describes the path between the device, home network, broadband connection and test service at that moment. It is not the same as a controlled laboratory certification of an internet provider.
Core rule: repeated wired and wireless tests are more useful than one isolated result. Provider rankings and local claims are not based on a single test, an invented score or an undisclosed star rating.
Broadband speed test methodology
The main LinkSpeed test runs in the visitor's browser and sends test traffic directly to Cloudflare's speed-test service. The LinkSpeed web server serves the page and receives a small anonymous result summary; it does not proxy the large download and upload payloads.
Upload speed
Estimates how quickly the browser sends test data. Background backups and asymmetric broadband packages can make upload results especially variable.
Ping
Measures browser request round-trip time to the test endpoint. It is useful for responsiveness but is not the same as ping to a particular game server.
Jitter
Summarises variation between latency samples. Lower and steadier figures usually indicate a more consistent connection.
For a cleaner line baseline, pause downloads, disable VPNs and test over Ethernet. Repeat over Wi-Fi to measure the experience of the device and room actually used.
Ping, packet-loss, bufferbloat and loaded-latency tests
The specialist tools separate different symptoms rather than treating download speed as the only quality measure.
- Ping test: samples HTTPS request latency and failed requests to the fixed LinkSpeed endpoint. Completed summaries can contribute to a separate specialist ping dataset.
- Packet-loss test: reports failed or excessively delayed browser test requests; it is not a raw ICMP network-engineering capture. Completed summaries can contribute to a separate browser-request reliability dataset.
- Loaded latency: compares quiet-line latency with latency measured while download and upload traffic load the connection. Completed summaries can contribute to the specialist loaded-latency dataset.
- Bufferbloat test: remains a diagnostic interpretation tool for dataset purposes. Its results are not stored separately because it measures substantially the same idle-versus-loaded behaviour as the canonical Loaded Latency Test.
- Interpretation: a poor wireless result followed by a good Ethernet result points first towards the local Wi-Fi path rather than proving an ISP fault.
Anonymous result storage and aggregate reporting
The public recent-results feed and main quality dataset use a restricted speed-test summary: test time, download, upload, ping, jitter, broad quality label, broad provider, country, city and device category. IP addresses are not written into that public result dataset.
Selected diagnostic tools also write a separate specialist summary. Depending on the tool, this can include dedicated ping and jitter, failed browser-request counts, connection type, loaded-latency medians, added latency, throughput, grade, broad provider, city and country code. The specialist store does not include the visitor's IP address or exact postcode.
The UK Broadband Quality Report keeps these families separate. Main speed-test provider rows require at least 10 qualifying tests and city rows at least five. Specialist provider comparisons require at least 20 Ethernet-labelled UK tests for that provider; Wi-Fi and unknown-connection results are excluded from specialist provider tables.
Results are self-selected and are not weighted to represent every UK household. Specialist tests are particularly likely to be run by people diagnosing a problem, so they are evidence about the LinkSpeed sample rather than a nationally representative ISP league table or a replacement for Ofcom's designed research.
Availability maps and postcode checks
Coverage tools use local copies of published Ofcom Connected Nations data and postcode-reference information, then combine that with clearly labelled provider and network links. A postcode or city result is contextual: final service availability, installation route and package choice must be confirmed at the exact address with the provider.
City pages are not generated as thousands of interchangeable doorway pages. Each guide is expected to explain local network overlap, flats and wayleaves, common infrastructure routes and the limits of city-wide statistics.
Provider comparisons and commercial guides
Provider and deal pages are checked against official provider websites, published contract terms, Ofcom material and LinkSpeed diagnostic evidence where a sufficient sample exists. Prices and promotions are date-stamped because they can change by postcode and expire quickly.
- LinkSpeed does not claim to buy and laboratory-test every broadband package.
- No AggregateRating or star score is published unless genuine visible user ratings and a defensible collection process exist.
- A low advertised monthly price is not treated as the cheapest deal until setup charges, scheduled rises, rewards and minimum-term cost are considered.
- Affiliate or commercial links should be disclosed and do not determine the technical diagnosis.
Editorial standards, reviews and corrections
- Clear intent: answer the user's problem before adding background detail.
- Primary evidence: prefer regulators, official provider terms and technical documentation for changeable factual claims.
- Visible limitations: distinguish measured data, provider claims and editorial interpretation.
- Material updates: change page and sitemap dates when primary text, structured data, links or methodology materially changes—not for a routine copyright rollover.
- Corrections: investigate reported errors and update or remove claims that cannot be supported.