Delay is high even while the connection is quiet
Repeat on the same device and path, then record the application and time rather than diagnosing a provider route.
Notice next: Capture the idle test, device, connection type and affected task.
High delay with strong capacityU.S. national decision guide
Internet delay decision guide
Reviewed and updated September 24, 2026
Why can latency be high when a speed test looks fast?
A fast download result and a responsive connection measure different dimensions. RFC 9318 describes capacity and packet latency separately: a connection can carry substantial data yet still make an interactive task wait for responses. The report also distinguishes idle latency from working latency, so a quiet-network result may not represent delay while uploads, downloads or other household activity are using the connection.
Compare the same device, connection path and task under quiet and busy conditions. Record whether the delay affects calls, games, browsing or every activity, and whether it changes on Ethernet versus Wi-Fi if an owner-accessible wired test is available. Do not turn one number into a diagnosis of the ISP or declare a universal acceptable latency for every application.
Safe test boundaryKeep testing to normal device and router controls; do not access provider or building equipment.
See what changes by location
What remains property-specific
Access technology, route and household load change the observation.
U.S. national guidance
What changes after you choose a location?
Choose a state to make the geographic boundary explicit.
A ZIP code does not prove price, coverage or job acceptance.
Applies nationallyThe decision structure is general, not a remote diagnosis.
Must be confirmed locallyProduct instructions and local requirements still control.
Separate safety, observed conditions and project scope
Which situation best matches the home?
Delay rises while the connection is busy
Record which upload, download or household activity is active when responsiveness changes.
Notice next: Compare quiet and working conditions without changing several variables at once.
The problem appears only on Wi-Fi
A safe Ethernet comparison can separate the local wireless path from the broader connection, but it does not name the cause.
Notice next: Record location, band, device and whether the wired comparison differs.
Safe observation
What to capture
Use comparable test conditions.
Measurements describe the symptom; they do not diagnose an ISP.
What to record
- Idle and working latency
- Device and Wi-Fi or Ethernet path
- Affected application
- Concurrent upload or download activity
- Time and repeatability
Stop and get appropriate help
- Do not open network equipment or outdoor enclosures.
- Do not interrupt another person's critical connection without agreement.
- Stop a test that overheats equipment or requires unsafe cable access.
Short answers
Questions people ask before the call
Does a fast download result rule out latency?
No. Capacity and packet delay are separate network-quality dimensions.
Which latency number is acceptable?
No universal value is supported here; application needs and test conditions differ.
Why compare idle and working conditions?
RFC 9318 distinguishes them, and household traffic may change the observed delay.
Sources and boundaries
What supports this guide?
- Internet Architecture Board / RFC Editor — RFC 9318: Measuring Network Quality for End-UsersCapacity versus latency and idle versus working latency; no service guarantee. Scope: national. Source checked 2026-09-24. Confirm address-specific requirements with the relevant local authority or qualified professional.
RFC 9318 is an informational workshop report.
No ISP is diagnosed.
No universal acceptable value is asserted.
A speed test is not proof of service quality.