How to run an accurate broadband speed test

An accurate test removes self-inflicted noise: use a wired connection when you can, pause competing traffic, pick a time that matches the problem you care about, and average several runs. Treat one flashy number as entertainment, not evidence.

Quick answer

Most important habitSame device, same room, same hour - then change one variable at a time.
When to retestAfter router moves, new mesh nodes, or when your provider says they fixed something.
What Pulse can and cannot confirmDownload, latency, and jitter in-browser - not upload; see upload scope.

Numbered method

  1. Connect with Ethernet if possible and close heavy apps, updates, and cloud sync on that device.
  2. Pick the time window you care about - quiet baseline or busy-hour realism - and stick to it for a set of runs.
  3. Open Pulse, run the test three times in a row, and jot each Mbps, latency, and jitter.
  4. Repeat the trio on Wi-Fi in the room where you actually work or play, same hour if you can.
  5. Compare medians: large Ethernet versus Wi-Fi gaps usually mean in-home wireless work, not the exchange.

Mistakes that skew results

Typical home scenario

A user in Norwich wants to complain about “slow fibre.” Their phone on kitchen Wi-Fi shows 38 Mbps; the same hour, a wired desktop shows 68 Mbps. That 30 Mbps gap is local - not proof the ISP failed - until a fair test is logged.

How Pulse relates to this topic

Pulse is built for quick, repeatable reads of download, latency, and jitter without storing your results server-side. Follow methodology so your setup matches what the numbers mean.

Once you have a believable download Mbps, optional download time and household speed helpers on the main site can translate numbers into rough time or range estimates - they do not replace testing your line.

Run the Pulse speed test · Read methodology · Review privacy

FAQ

Should I test on Wi-Fi or Ethernet first?

Ethernet first if you can. It isolates the broadband line from wireless quirks; add a Wi-Fi run to see the gap.

How many runs are enough?

At least three back-to-back on the same setup, then repeat on another evening if you are chasing a pattern.

Does VPN affect Pulse?

Often yes. Turn VPN off for a baseline unless you deliberately want to measure through it.

Can I test while the family streams?

Yes, if you want a realistic busy-house number - but label it as such; it is not your line maximum.

Sources and review notes

Last reviewed: 11/04/2026 · Written by: Dr Alex J Martin-Smith (LinkedIn)

Related guides