Advertised Internet Speed vs. Your Test: A Practical Checklist

Establish what your plan actually says

Save the plan document supplied for your service, not a current advertisement for another tier. In the US, the FCC's Broadband Consumer Labels guidance describes labels covering prices, introductory rates, allowances and speeds. Follow the fields on the actual offer and ask the provider to clarify anything missing.

Keep download and upload information separate. Record the service address privately, the plan identifier, any equipment supplied and the date of the document. A general national comparison cannot tell you which tier is provisioned at your home.

Build a controlled baseline

Ofcom recommends taking tests over several days at different times, and trying Ethernet rather than wireless. Its practical speed guidance identifies factors including household use, device setup and time of day. Use those checks to narrow possibilities, not to assume a particular provider is at fault.

1. Use the same computer and test service where practical. Record the endpoint if the tool exposes it.
2. Connect to the router by Ethernet for a baseline. Record the connection method and any adapter in use.
3. Pause discretionary transfers. Leave required security and work controls in place, and note a managed VPN rather than bypassing it.
4. Save download, upload and available latency information, plus the date, time zone and whether the test completed. Keep errors visible.
5. Repeat at other times without changing the whole setup. Then test the real Wi-Fi location as a separate condition.

A test itself transfers data. Consider your plan's data allowance before repeating large tests, especially on a metered connection. You do not need a long stress test to start a useful support conversation.

Copyable blank record: separate the promise from the observation

Copy this template into a private note. The brackets are blanks, not collected results. For missing plan terms write not supplied; for an unsuccessful run record the failure and any warning, not a zero-speed result. Do not put an account number, full service address or public IP address into a public post.

PLAN DOCUMENT
Exact plan identifier: [copy from your own offer]
Document date and source link: [record both]
Download wording and units: [quote the stated term; Mbps if supplied]
Upload wording and units: [quote separately; do not assume symmetry]
Typical / advertised / minimum: [retain the actual distinction]
Missing terms to ask about: [not supplied]

ONE OBSERVATION — repeat this block for each attempt
Date, local time and timezone: [fill in]
Device / OS / browser: [fill in]
Connection: [Ethernet or Wi-Fi; adapter and room]
Test service and endpoint: [record; unknown if not exposed]
VPN / required controls / other traffic: [record conditions]
Download Mbps and upload Mbps: [reported values, or unavailable]
Latency metric and units: [name the tool's actual measure]
Completion / failure / validity warning: [retain the exact message]
Affected application and symptom: [describe separately]
Change since the previous attempt: [one change, or none]

PRIVATE SUPPORT REQUEST
Please verify the provisioned tier, connection faults and supplied equipment.
Case reference and agreed next step: [fill in after support responds]

The FCC's label guidance explains where US plan price, allowance and speed disclosures belong. It does not make every stated speed a contractual minimum. Save the exact document and ask the provider which term applies to your service. For alternatives, use the blank offer-comparison worksheet.

Keep Mbps separate from MB/s, download separate from upload, and the plan description separate from a particular test. If the endpoint, device or interface changed, mark that attempt as a different condition. If only one application fails, include its own symptom rather than treating a general speed test as its performance measurement.

Interpret patterns without overclaiming

Wired results are consistently better than the problem Wi-Fi location

Investigate placement, coverage, interference and the client device. This pattern points toward a local wireless constraint; it does not prove the access service is fault-free.

Both wired and wireless tests are poor

Check whether the device, adapter or router could be limiting the test, and ask the provider to check the provisioned tier and connection. Do not jump directly to a technology-wide conclusion.

Latency rises mainly during a transfer

Queueing under load is a possibility. OpenWrt's SQM documentation explains how queue management can mitigate this kind of delay, with hardware and configuration limits.

One application is slow while general tests look normal

Record the application, selected region or destination, time and error. The speed-test path is not necessarily the path used by that service.

Repeated poor evening results can justify further investigation, but timing alone does not prove cable-node congestion, throttling or deliberate prioritisation. Keep those as hypotheses unless more specific evidence establishes them.

Send support a concise, reproducible summary

Use this record in a private support channel:

• Plan: exact identifier and stated download/upload information.
• Symptom: slow transfer, disconnects, or delay while other traffic is active.
• Setup: device, wired/Wi-Fi, router, relevant software and unavoidable background use.
• Observations: dated results and errors, with endpoint changes identified.
• Checks already tried: what changed and whether it helped.
• Request: check provisioning, line faults and supplied equipment; agree on a next step and obtain a case reference.

Redact account identifiers, addresses and other personal information before sharing a record publicly. Cancellation rights, minimum-speed commitments and remedies depend on the contract and jurisdiction; this checklist does not establish an entitlement.

What these observations cannot establish

Your tests can document an experience and help narrow causes. They do not estimate a national median, prove how all customers of a provider perform, or justify a brand-wide score. We are not presenting a field-study population, delivery-ratio ranking or rolling aggregate in this guide.

For comparing an offer with alternatives, use the unranked ISP comparison guide. For spotting misleading population claims and statistics, see how to read broadband speed reports. If hardware is the suspected constraint, use the router selection checklist before spending money.