Complete Guide to Starlink Speed Tests

A speed test result is a single number pretending to be a verdict. You run it, see a strong result, and either smile or start a support ticket. But that one number hides four different things happening at once, and the real question is whether your Starlink is performing normally for your plan, in your location, at this hour. A modest reading in a remote Alaskan cabin at 9pm may be spectacular. The same result in downtown Austin at noon on Residential is a red flag.

This guide walks through what your test actually measures, what to expect by region and plan, how to run a test that isn't lying to you, and the thresholds where support becomes worth your time. If you want to skip the theory and get a baseline right now, our free in-browser speed test measures download, upload, ping and jitter against servers on three continents and logs your history so next week's test actually means something.

What a speed test actually measures

Every legitimate speed test reports four numbers. Most people only look at the first one. Here's what each one really tells you:

Download speed (Mbps) measures how fast a big chunk of data can come toyou. This is the headline number, and it's what matters for streaming Netflix, loading web pages, downloading game updates, and almost every consumer use. A single 4K Netflix stream needs a stable share of the connection. A family with three devices streaming at once needs more headroom. Beyond a comfortably fast connection, additional download capacity is essentially indistinguishable for typical home use — the bottleneck is your device, not your pipe.

Upload speed (Mbps)measures the reverse direction — you sending data out. This matters for video calls (Zoom uploads your webcam), cloud backups, uploading photos, and live streaming. Starlink's upload is structurally slower than its download because the spectrum allocation is asymmetric. A modest, steady upload is enough for HD video calls; stronger upstream capacity is comfortable for simultaneous calls and backups; sustained low upload is where calls start freezing.

Ping / latency (ms)is the round-trip time for a tiny packet to reach a server and come back. This is not about bandwidth — it's about responsiveness. Low ping matters for video calls, gaming, and anything interactive. Starlink's low-Earth orbit reduces propagation delay; a typical Residential ping still varies by route, load, and handoffs. Traditional geostationary satellite (HughesNet, Viasat) has much higher ping, which is why video calls over those services are miserable.

Jitter (ms) is the variation between consecutive ping samples. A steady result is preferable. A wild latency swing — even if the average is low — will destroy a voice call. Jitter is the best early-warning signal that a cell is congested or a dish is losing satellite handoffs. Healthy Starlink jitter stays low and stable.

Typical Starlink speeds by region and plan

Read these rows qualitatively because results vary by location, plan, cell load, test method, and time of day. Use Ethernet-wired tests to separate the satellite link from local WiFi loss.

If your numbers land inside these bands, Starlink is working normally for your location. If they don't, the next sections tell you where to look. To see exactly where your results fall against your specific region and plan, plug them into our ISP comparison tool which benchmarks against 20+ alternatives.

Why your speed varies hour to hour

Starlink is a fundamentally different beast from fiber or cable. With fiber, you get a dedicated pipe and your speed is essentially constant. With Starlink, you share a radio cell with hundreds or thousands of other subscribers, and the satellite overhead changes every few minutes. Four things drive most of the variation you'll see on any given day:

Satellite handoff and pass geometry. Starlink satellites are in low Earth orbit — they move. Every 2–4 minutes your dish hands off to a new satellite. During the handoff, you may see a 1–3 second latency spike or a brief bandwidth dip. If one handoff glitches, your next speed test will show it. Run the test three times in a row and take the best number.

Obstructions.Even partial sky blockage — a branch that grew over summer, a new building, a flagpole — can drop your dish's effective uptime from 99.9% to 95% and hammer your test results. The Starlink app shows an obstruction map, and our obstruction planner lets you simulate different mount heights and positions before you drill holes. Fixing obstructions is the single highest-leverage improvement for a slow dish.

Time of day and cell congestion. Every Starlink cell has a finite capacity. Between 7pm and 11pm local time, demand peaks — kids home from school, parents streaming, gamers online. A cell that delivers strong throughput at noon can slow sharply by 9pm. This isn't a fault; it's the design. If you run a speed test only at 8pm and panic about the number, you're sampling the worst-case window. Always test at least once at noon to know your ceiling.

Weather.Heavy rain, snow accumulation, temperature, and movement of the installation can affect performance. The impact is not a universal loss percentage or temperature threshold across every terminal. Compare clear-weather and adverse-weather tests, and consult Starlink's official specifications for the exact model.

How to run a proper test

If you're going to complain to Starlink support, or make a keep-or-cancel decision, you need numbers you trust. Here's the methodology that removes most of the noise:

1. Wired, not WiFi.Plug directly into the Starlink router via the Ethernet adapter, or use a third-party router in bridge mode. WiFi introduces so much variability (interference, client NIC, wall penetration) that it masks what's actually arriving at your dish. Test wired first to know the ceiling, then test WiFi to measure how much your home network is costing you.

2. Quiet the network. Pause cloud backups, stop streams, close browser tabs, and make sure no one is downloading game updates. A single background 4K stream will consume part of your test bandwidth and make the result look lower than the idle-network baseline.

3. Reboot the router and dish.If it's been a week since the last reboot, power-cycle both and wait 5 minutes for reacquisition. This clears accumulated routing-table cruft and forces a fresh satellite handoff.

4. Run three tests, take the median. Back-to-back tests catch transient handoff dips. Three tests in a five-minute window gives you a median that filters out one-off glitches.

5. Test at three times of day. Morning (8-10am), mid-afternoon (1-3pm), and evening peak (8-10pm). The spread between these three tests tells you more about your connection than any single number. A healthy Residential connection may slow only in the evening; that pattern points to congestion. A troubled connection stays slow even off-peak.

6. Log your results. A single test tells you nothing; a month of weekly baselines tells you everything. Our speed test stores your history locally in the browser so you can watch trends over time instead of reacting to noise.

How to read your result in five checks

Once you have a clean wired result, read it in this order. Each check rules out one common misreading before you blame the satellite link.

1. Check the unit. Speed tests report megabits per second (Mbps). File-download windows usually report megabytes per second (MB/s). One byte is eight bits, so you divide Mbps by 8 to get MB/s.

The times assume 1 GB is 8,000 megabits and the link is fully used, so real downloads run a little slower because of protocol overhead and the server on the other end.

2. Treat the plan speed as a ceiling.A plan's headline speed or published range is the best a connection should reach, not a promise for every hour. Starlink lists its current typical ranges per plan on its specifications page, and plans and ranges change, so read them there rather than from a table in an article. Then measure your own best off-peak wired result and use that as your baseline. Your later tests are judged against that baseline, as a share of it.

3. Match the number to what you do. A result is good when it covers the activities running at the same time with room to spare. Netflix, for example, recommends 3 Mbps for 720p, 5 Mbps for 1080p and 15 Mbps for 4K per stream. As an assumed example, a household running two 4K streams and one 1080p stream needs 15 + 15 + 5 = 35 Mbps for video alone, before a video call, a game download or a phone backup. If your busiest-hour result sits close to that total, you will see buffering even though the test number looks large.

4. Read download, upload, ping and jitter by symptom. Start with the number that matches what you actually notice.

5. Compare like with like. A different server, a different number of connections, a VPN or a browser extension all change the result. Two tools can both be right and still disagree, so keep one tool and one method and compare it with itself. Our methodology page explains how SatSpeedCheck measures, including what it reports as not measured rather than as zero.

Worked example with assumed numbers

Assume a wired off-peak baseline of 180 Mbps down, 18 Mbps up and 35 ms ping. At the evening peak the same wired test gives 95 Mbps down, 12 Mbps up and 48 ms. The download is 95 / 180 = 53% of baseline, a drop of about 47%, and ping is higher but not wild. That pattern, slower in the evening and normal off-peak, points to congestion rather than a fault, so a support ticket is not the first step.

Now assume the same 180 Mbps wired result, but 60 Mbps on Wi-Fi in the back bedroom. That is 60 / 180 = 33% of the wired speed, and the satellite link is not the bottleneck, because the wired number is healthy. Move the router, switch bands or add a wired run before changing anything on the dish. If instead the wired result itself stays far below baseline at noon with a clear sky, that is the case worth diagnosing: see why your Starlink is slow, the ping and latency guide, and jitter explained. For the gap between a plan's headline and your real results, read actual vs advertised speed, and for the wired setup used above, the Ethernet setup guide.

Starlink speeds vs the alternatives

A healthy Starlink connection looks slow next to gigabit fiber and looks like magic next to HughesNet. Context matters. Here's how Starlink Residential stacks up against the realistic alternatives most rural and semi-rural users are actually comparing:

The story the table tells: Starlink can outperform legacy rural options, while fiber generally wins on raw speed and latency. Exact results depend on the address and provider. Walk through your specific scenario in the ISP comparison tool to see which alternative — if any — would serve you better at your address.

When to actually worry (and what to do about it)

Most slow speed tests aren't a Starlink problem — they're a WiFi problem, a time-of-day problem, or an expectation problem. But some results genuinely warrant action. Use these thresholds:

Wired Residential download stays well below its off-peak baseline with clear sky: this is a real issue. Either you have an unreported obstruction, your dish firmware needs a reset, or your cell is oversubscribed beyond what Starlink should allow. File a support ticket with your timestamped speed-test screenshots.

Ping unusually high sustained with high jitter: your dish is likely struggling with satellite handoffs. Check the obstruction map in the Starlink app and run our obstruction simulator to check whether a different mount location would fix it. Most of these cases are solved by raising the dish 4–8 feet or moving away from a tree line.

Frequent dropouts (connection lost for 10+ seconds, multiple times per hour):check the Starlink app's outage log. If it attributes the drops to obstructions, fix the view. If it attributes them to “no satellites,” you're in an edge-of-coverage zone and may need to wait for more satellites to deploy — or accept the reality. If outages are unattributed, try a dish reboot and a router-bypass test.

Upload consistently with modest throughput: this one is almost always a cell-congestion issue on Residential, since the uplink is the first thing throttled under load. Test upload at 3am; if it improves substantially, congestion is confirmed. Priority plan is the only real fix.

Before contacting support, also check our total cost calculator to make sure the plan you're on matches what you actually need — some of what people call “slow Starlink” is really “wrong plan for my usage.”