Your internet runs fine at 9 a.m. Then 8 p.m. hits, Netflix starts buffering, and your ping in-game spikes past 300ms. You restart the router. You restart the modem. Still slow. That pattern is not random. It has a name: throttling. And your ISP is almost certainly the one doing it.
Throttling Diagnosis at a Glance- Run speed tests at off-peak and peak hours and compare the results
- Measure your real speeds against your plan's advertised rates
- Test streaming, gaming, and large downloads separately to catch selective throttling
- Run a second speed test with IP masking active to expose traffic-based slowdowns
- Log every result with timestamps before contacting your ISP or filing a complaint
Why ISPs Throttle Traffic in the First Place
ISPs have the technical ability to inspect and shape traffic flowing through their networks. This practice, known as bandwidth throttling, gives them a way to manage congestion and prioritize certain data flows. In theory, it stabilizes the network for everyone during heavy use. In practice, it often means your Netflix stream gets deprioritized while your ISP's own streaming service gets a free lane.
The FCC's Measuring Broadband America research has tracked real-world ISP performance for years and consistently found gaps between advertised speeds and delivered speeds. Throttling is one driver of that gap. Not every slowdown is throttling, though. A congested Wi-Fi channel, a failing modem, or a corroded coaxial line can produce identical symptoms. The steps below help you separate the causes.
Building a Baseline with Timed Speed Tests
A single speed test is almost useless on its own. You need data across time. The goal is to identify whether your speeds are consistently low or only drop at predictable intervals.
Start early in the morning, between 6 a.m. and 8 a.m. Network traffic is low at this hour. Your ISP has little incentive to throttle, and congestion is minimal. Record your download speed, upload speed, and ping. That number becomes your baseline.
Then run the same test during peak hours: 7 p.m. to 11 p.m. on a weekday. This is when your neighborhood floods the shared network. If your speeds drop sharply during this window and recover overnight, that pattern points directly at congestion-based throttling.
Repeat this over four to five days. A single bad evening could be a temporary outage. A consistent pattern of evening slowdowns is much harder to explain as coincidence.
Comparing Your Measured Speeds Against Your Plan
Pull up your current plan details. Xfinity, Comcast, and AT&T all list advertised speeds on your bill and in your account dashboard. These are the numbers your ISP is selling you.
A healthy connection delivers between 80% and 95% of advertised speed on a wired device during off-peak hours. Anything under 50% of advertised speed at a low-traffic time is a problem, but it may not be throttling. A failing modem or a degraded cable line can produce the same result.
An Xfinity customer on a 400 Mbps plan should see 320 to 380 Mbps on a hardwired connection during off-peak hours. An AT&T Fiber subscriber on a 1 Gig plan should land close to 940 Mbps. If your baseline numbers are nowhere near those marks, start by testing the hardware before assuming the ISP is throttling you.
Testing Specific Traffic to Catch Selective Throttling
Standard speed tests use their own data protocol. ISPs often leave speed test traffic untouched because it is visible and measurable. That means a speed test can show healthy numbers while Netflix, gaming servers, and large downloads are getting quietly strangled.
Testing specific traffic types catches this:
- Stream a 4K title on Netflix or YouTube for 10 minutes and note how often quality drops or buffers
- Start a large game download through Steam or a console platform and watch the sustained transfer rate over several minutes
- Join an active multiplayer session and track ping and packet loss inside the game client
- Download a large public file via direct HTTP link and record the average speed throughout
If your speed test reports 250 Mbps but Netflix degrades to 480p at 9 p.m., that gap is a red flag. Streaming traffic is especially easy for ISPs to identify by destination IP address and traffic signature. Targeting it lets them reduce costs without showing up as a blanket speed reduction.
Using IP Masking to Confirm Traffic-Based Throttling
This is where the diagnosis gets precise. When your ISP can see your real IP address, they can fingerprint your traffic type. They know you are streaming to a Netflix server. They know you are connecting to a PlayStation gaming node. They can throttle those specific flows while leaving general browsing untouched.
Hiding your IP disrupts that fingerprinting. Once your ISP cannot identify the nature of your traffic, they have no basis to selectively deprioritize it.
Run a standard speed test and note the result. Then activate IP masking and run the identical test again. A meaningful speed increase with masking active is strong evidence that your ISP was throttling based on traffic identification. With your fingerprint hidden, the throttling had no target.
Pair this with your traffic-type tests for more detail. Try streaming Netflix or running a game download with masking enabled. If those activities perform significantly better than they did without it, you have narrowed the cause. Your ISP was identifying and deprioritizing your connection type, not applying a blanket speed cap.
Speed Test Patterns and Their Likely Causes
Different results point to different problems. Reading the pattern correctly saves time before you call your ISP.
What Each Pattern Usually Means
| Speed Test Pattern | Likely Cause | Recommended Next Step |
|---|---|---|
| Speeds drop only at peak hours | Network congestion on a shared node | Document timing and ask ISP about local node capacity |
| Speeds low at all hours | Hardware fault or degraded line | Test wired directly to modem, request a line inspection |
| Speeds improve with IP masking active | Traffic-based throttling by ISP | Build a documented test log and escalate formally |
| Specific services slow but speed test looks fine | Selective throttling by destination | Log affected services and peak windows, file FCC complaint if unresolved |
What Xfinity, Comcast, and AT&T Users Should Know
Xfinity and Comcast operate on the same hybrid fiber-coaxial network. In densely populated areas, multiple households share a local node. That shared infrastructure makes peak-hour slowdowns common, and it also makes selective throttling difficult to distinguish from genuine congestion. Xfinity subscribers have reported throttling of Netflix and Amazon Prime Video traffic during evening hours across several markets. The diagnostic steps above are particularly useful for Xfinity customers because the congestion excuse is always available as cover.
AT&T Fiber delivers more consistent performance because fiber connections are not shared the same way coaxial nodes are. A dedicated fiber line means fewer variables between you and the ISP's backbone. However, AT&T's legacy copper DSL product and its fixed wireless Internet Air service behave much more like cable in terms of peak-hour variation. If you are on one of those plans, apply the same diagnostic approach.
Comcast carries a documented history with net neutrality enforcement. The FCC ruled in 2008 that Comcast had illegally throttled BitTorrent traffic, a decision that shaped regulatory debate for years. The technical capability for traffic-based throttling has not diminished since then. It has only become harder to detect.
Building a Log That Your ISP Cannot Dismiss
If you call Xfinity or AT&T support and say your internet feels slow, a technician will run a line test and close the ticket. If you arrive with a week of time-stamped speed results, a comparison to your advertised plan speeds, and a list of affected services, the conversation is different.
Keep a simple log for each test. Note the date, time, download speed, upload speed, ping, what activity you were testing, and whether IP masking was active. Seven days of data across peak and off-peak hours gives you a credible, hard-to-dismiss record.
When contacting your ISP, ask them to explain the gap between your tested and advertised speeds. Ask specifically whether traffic management policies apply to your tier. Request written confirmation of any policy they describe. Keep records of every call including agent names and ticket numbers.
If your ISP does not resolve the issue, file a complaint at fcc.gov/consumers/guides/filing-informal-complaint. The FCC forwards complaints to the provider and requires a formal response. That added pressure tends to produce faster results than support queues alone.
Turning Test Results into Real Answers
One slow evening is noise. A consistent pattern of peak-hour drops, speeds that climb when masking is active, and specific services crawling while your general speed test reads healthy: those add up to signal. No single data point proves throttling. The case is built by combining tests.
Start with your morning baseline. Measure it against your advertised plan. Test the traffic types that matter to you: streaming, gaming, large downloads. Apply masking to isolate traffic-based throttling. Write down every result with a timestamp. Then escalate with evidence rather than frustration. That approach gives you the clearest possible picture of what your ISP is delivering versus what you are actually paying for.