Thank you for being a valued part of the CNET community. As of December 1, 2020, the forums are in read-only format. In early 2021, CNET Forums will no longer be available. We are grateful for the participation and advice you have provided to one another over the years.

Thanks,

CNET Support

General discussion

Are internet speed connection tests rigged?

Aug 5, 2016 5:50PM PDT

I have a question regarding internet connection speed. Are the so-called "speed checks" rigged? I ask this because recently I received an email from CNET that contained a solicitation inviting you to check your connection speed. Well, I did and my result averaged around 6Mbps. My ISP is AT&T, which boasts its U-Verse is superior. I'm subscribed to their 12Mbps plan. So after I learned that the solicitation link clocked me at 6, I tested my connection on U-Verse's speed check which shows me clocked at around 12Mbps download and around 3 upload, which is fine but makes me wonder which is right? I've always kind of wondered what the actual speed is when at times I find myself waiting patiently staring at a screen as the page loads. Any thoughts or previous discussions on this subject? I'm wondering if speed is just a marketing scheme to entice one believing their connection is better than their competitor. And is there a trustworthy method of finding out my actual connection speed without bias? Thanks.

--Submitted by Alan L. of Columbus, Ohio

Post was last edited on August 5, 2016 5:52 PM PDT

Discussion is locked

- Collapse -
your speed test and distance
Aug 5, 2016 6:24PM PDT

I honestly believe some of the tests are rigged. at the same time, the distance to the server that is doing the test must also be taken into account. In addition you must take into account the traffic between you and the testing server.

I will never use the isp's speed test link. I will use speedtest's link though it could have have been given priority status by the isp. the reality is, I am a cord cutter and as long as there is no buffering of the stream, it really does not matter.

- Collapse -
I agree
Aug 15, 2016 4:23AM PDT

The distance, traffic and the server are very important factors and they will vary by the time of day. Another thing to take into account is how the ISP is testing. They may only be testing between their core network and your modem/gateway installed at your house. The 3rd party test will cover the ISP network, plus the Internet hops between your client computer and their server.

No matter which test you use though, one thing to note is YOUR connection. If you are using a laptop over Wi-Fi, that will reduce your speed results. I check my connection from a wired computer connected to the modem as well as from a Wi-Fi client. My ISP says that I have a guaranteed 75Mbps download and I can get up to 92Mbps on the Ethernet connected client and about 20Mbps (+/-) on a good day over Wi-Fi using speedtest.net.

- Collapse -
ISP can favor their own tests...
Aug 5, 2016 6:47PM PDT

Use DSLreports HTML5 tests with options to pick servers (i think you can pick up to 5)

That will give you an accurate test. In USA, fhere are net neutrality laws against throttling, but isps in other parts of the world likely do it for bittorrent or other traffic that might hog bandwidth.

Ive heard of people using seedbox and then do direct http/ftp download from seedbox

- Collapse -
RE: ISP can favor their own tests...
Aug 13, 2016 3:44AM PDT

I agree about DSL Reports test being good, but it's also just a picture in time. Best options would be to graph yourself to see how things vary over time. The most well known project in the US would be SamKnows: https://www.samknows.com/. They will give you a router and allow you to see how well the ISP keeps it's speed consistently. Data is then also shared with regional regulators like the FCC to ensure that consumers are not being ripped off. Like you stated, you could do the tests yourself, with a OpenWRT router, linux box, et al, but I would suggest using more than 1 location as their could be external issues like a fiber cut, BGP loop/leak, etc that could limit speed at times to a location but not others.

- Collapse -
Your Internet Provider Gives the "Best Case" Speed Test
Aug 5, 2016 7:34PM PDT

AT&T or any other provider will connect you to their nearest server that isn't bogged down with heavy traffic. That way you get a "fast" result. To get an idea of how much variation there is just on big servers just try Speedtest.net. The first time you use it the server nearest you will probably be used for a fast result. Then, you can choose a different location. Try somewhere a long way off or maybe another country and you'll see how bouncing your test request from city to city causes some slowdown. And, any website you visit while surfing the internet will have fast or slow or very busy servers that enhance or limit the page loading speed, downloads, etc. In other words, the speed you pay for is only the maximum you may get under ideal conditions.

- Collapse -
Speed Tests
Aug 5, 2016 8:05PM PDT

If you are using a wifi connection there is a serious drop-off in speed versus a direct wired connection. Most ISPs advertise and deliver speeds based on a wired connection. In my experience ookla's speedtest.net reports fairly accurate speeds and you can choose the server location for the test.

- Collapse -
Slow WiFi
Aug 12, 2016 8:39PM PDT

This is mostly true. However, I did a little research and found that COX will provide a high speed wifi router but you must ask for the particular router specifically. It is an Arris TG2472. This router gets me 400 Mbs wifi. Modem rental is $6.99 a month.

So if you are getting slow wifi then you need a new modem.

- Collapse -
Speed or bandwidth?
Aug 5, 2016 8:10PM PDT

Don't forget that speed and bandwidth are two different things. A site could be getting a lot of traffic which can affect your page loading or maybe the time of day/night in your area is particularly heavy.

- Collapse -
Speed test reliability
Aug 5, 2016 9:06PM PDT

As I read your inquiry, I was laughing to myself, about the many times I had the same question. When I bought into the DSL, 12mbps, through my phone company, I was expecting a definite change in how fast things would upload. Of the many times I complained to the company, because of very slow speeds, I was always given the excuse, my computer was needing to be cleaned out. Also things would go from 6mbps to 12mbps as soon as the tech "fixed" the problem, outside some where. Now I am getting some "New" high speed upgrade, on a free trial basis, for a year. I have again lost the 12 and fell to 6 again. I complained loudly, this time. They told me I had to use their own speed test. So I d/loaded it and WOW, I had 50mbps!
I asked the tech as to why I had two different readings, and he said, my phone company has gone to an experimental technology, and their speed test is accurate, and the internet ones can't read theirs properly. Has anyone heard about this new technology? Can it work through the same wiring that I have talked on the phone for years? Am I getting duped?
I have a two year old HP 500-214 desktop, with 8mg RAM and 2T drive. I have Win10 that replaced the Win8.1 that was on it. The HP is hardwired into the Modem.

- Collapse -
Special method-like a Volkswagon smog test
Aug 5, 2016 10:55PM PDT

They are just throttling you down, and lying with their rigged test. If you believe them, go buy a used VW; the are clean burners, fuel efficient and powerful.

- Collapse -
Not just VW
Aug 12, 2016 5:41PM PDT

There's probably a better analogy. Dig deeper into the emissions ordeal. They were actually up against rigged rules.

- Collapse -
Re: Speed test reliability
Aug 15, 2016 4:28AM PDT

What they gave you with their "own speed test" that returned 50Mbps is an old technology wrapped up in a shiny new wrapper -- SNAKE OIL!

- Collapse -
I do believe your ISP is passing you taffy.
Aug 15, 2016 9:27AM PDT

Although your speed will vary their "explaination" sounds fishy.

- Collapse -
Ookla
Aug 6, 2016 6:32AM PDT

Go to www.speedtest.net. This is run by Ookla as a service and is not affiliated with any ISP. You'll get a fair reading.

(I have Verizon FiOS and pay for 75/75. I test it periodically and usually get results similar to what I just got: 83.42 down -- I have gotten as high as 89 -- and 86.64 up -- my high was 88.)

- Collapse -
ISP preferential treatment
Aug 12, 2016 5:19PM PDT

Just because speediest.net isn't affiliated with your ISP doesn't mean the ISP couldn't set it up that traffic to there servers get preferential treatment. Meaning you could still be getting a false reading.

- Collapse -
Response to testing
Aug 13, 2016 4:07AM PDT

Thank you for posting.....I just tested mine with speedtest.net --download 83.95 mbps upload 86.73 mbps
Then I tried Frontier (used to be Verizon in my area) 84.15 mbps download upload 86.14 mbps
I then went to CNet testing 6.7 mbps

Yes I am hooked up with Frontier (f/k/a Verizon in my area) at 75/75

So I guess I am happy with the results from speedtest.net which shows Frontier's test is ok.

- Collapse -
CNET's testing site
Aug 16, 2016 7:32AM PDT

like you i have tested my system with 4 different sites and got an average of 18 mbs down and 1-1.5 up. that is what i have with ATT on their Max Plus plan. I'm good with that. But CNET site gives me 6 mbs every time. What's to believe??

- Collapse -
How long is a piece of string?
Aug 6, 2016 7:13AM PDT

The only accurate answer to that question is twice as long as half of it! And that is pretty much true of DSL/ADSL speed. The main factors that will affect the result are the maximum speed offered on the line (as in the sales brochure, with or without the prefix "up to"), your distance from the exchange, the congestion on the line (consumer grade ADSL typically multiplexes 50 users per line) and congestion on the exchange.

Here in Australia, I have a theoretical maximum 20 Mbits/second on ADSL2+. We are 3.9 km from the exchange on copper wire. The actual maximum we get is 6 Mbits/second at quiet periods immediately after the modem has negotiated the connection with the exchange server. This on a wired Ethernet connection to the modem.

But we don't get this speed all the time. Generally, living in a fairly rural location, we don't experience too much line congestion, except when the schools finish for the day and the students are instantly on their mobile devices, or when there is a major sporting event being streamed by Foxtel (pay TV). We do, however, get serious exchange congestion on a regular basis.

Earlier this year, our download speed dropped to as low as 1.5 Kbits/sec - yes Kilobits! The ISP had switched our line to a hybrid (different logon credentials storage) and my modem was not fully compatible. The ISP supplied a new modem/router and that did improve the situation to around 5.5 Mbits/second. But over a week or so, it had degraded back to 2 Mbits/sec. Experimenting with various connections, it became clear that the download speed was being degraded at periods of heavy exchange congestion but the modem was not renegotiating with the exchange unless the line was pulled from the wall socket and reconnected. Switching the modem off and on didn't result in a faster speed, I concluded because when it was switched off, it was happy with the packet drop rate and so when switched on again started at that setting and was still happy.

The ISP replaced the modem/router again, this time with a later model and this one does periodically renegotiate the speed with the exchange so our speed range has stabilized to between 3.75 and 5.5 Mbits/second.

Just some of the factors that can play on the results. What I did notice when the ISP came out to check repeated complaints was that they always put their test equipment on the line before they ran their speed tests. They and I use Speedtest.net from Ookla and the tests pick a server from about half a dozen based in Brisbane, 25 miles from my exchange, based, they say, on the best ping.

When I lived in the UK, we had naked cable broadband (Internet only, no TV, no voice line) and that technology is inherently superior to DSL. I wish I could get it here! The UK cable ISP offers up to 400 Mbits/sec now for about the same price I pay here.

To get back to your question, are the tests rigged? They sure can be, so try to use an independent test facility.

- Collapse -
It Is Never Exactly the Same
Aug 6, 2016 1:19PM PDT

The way the Internet works, your connection to your destination (website/server) is rarely a direct connection. To show how this works, you need to use a traceroute utility. In Windows, for example, you can open a command prompt and type in "TRACERT www.cnet.com". You will see a number of HOPS. These are routers that grab your individual data packets, hold onto them and then send them off on a "path" toward your destination. You may notice that your computer retains this path information but it can change if there is a "blockage" en route. It could even be that to get to a server that is physically across the street from you, your packets could go by way of Russia.

So why would your ISP's test show better? Probably because the "target" server is topographically close to you and your packets don't actually go OUT onto the Internet but stay within the ISP's network. Geographic distance never matters. It is how many links or hops are between you and the destination (at any given moment) and the speed of each of those connections. Each "hop" determines which route it is going to use to send your packet based on "cost" which is actually an estimate of how slow each connection is, at that time. It is really fascinating looking at how TCP/IP really works. There is going to be a big difference between you communicating from your PC to your home router. Could be WiFi or a physical connection. After that, you use your ISP's network to get to your ISP's routers. All on their network and the speed you get there will vary with other traffic on that network, including the kid down the street watching an online movie! Once you leave your ISP's network, your packets can take any number of a different routes; some fast, some slow.

The other thing to watch out for is caching. You may notice to get to a new webpage that you never accessed before, it takes a little extra time but, after that, it seems faster. That is because your computer caches various elements of the page on your computer for reuse. Good speed testing software will bypass that cache but not all will do that. I suggest that when you run a speedtest, that you run it several times and, maybe, clear your browser's cache between each run. You will still see some variation. Also remember that your ISP's test target may be a lot closer to you on the network and not involve going out to the real Internet. Take what your ISPs says that you are paying for as a "maximum" bandwidth figure and you will be closer to the truth but you may even find the infrequent time when you will exceed that due to compression and other factors.

Post was last edited on August 12, 2016 1:32 PM PDT

- Collapse -
Wired results
Aug 6, 2016 9:41PM PDT

I just ran 5 tests; 3 using HTML and 2 (ookla powered) using flash. Here are the results for my wired connection.
speedof.me (HTML)
D: 72.93/U: 6.34

bandwithplace.com (HTML)
D: 82.16/U: 5.93

testmy.net (HTML)
D: 63.9/U: 3.4

speedtest.net (ookla)
D: 87.45/U: 5.95

speedtest.xfinity.com (ookla)
IPv4
D: 87.27/U: 5.98
IPv6
D: 85.94/U: 5.84

As you can see, 2 of the HTML based tests were significantly lower. My guess would be that they used different servers than the rest.
The HTML sites say they are more accurate because they don't use flash (which ookla does).
Comcast/XFINITY is the only one that gave separate readings for IPv4 and IPv6. As best as I have been able to learn, the others only test IPv4.

I also took the cnet speed test. The download speed was about half what Comcast showed. However, the upload speed was more than 10 times as high (over 40Mb/s). I don't know who they used, but I don't put much stock in their figures.

I recommend you use the first 4 listed above and compare the results. That should give you a fair idea what your actual speed is.
NOTE: Wi-Fi will be lower. The amount will depend on the throughput of your router and network card.

- Collapse -
Internet Speedtest
Aug 8, 2016 5:11AM PDT

I'm on FiOS and have used CNET's Speedtest for over a year and always received good results. When I was running 50 Mbps. My test result were around 56 Mbps. When I went to 75 Mbps, I got around 78 Mbps. However, ever since I installed Windows 10, I get 6 Mbps consistently. I get the exact same results on my two Laptops, but their on wireless. I also ran Verizon's and other speed tests. They all report around 75-78 Mbps. So, as it stands for now, it would seem in my eyes that the issue is an incompatibly between Windows 10 and CNET's Speedtest. Anyone else seeing this???

- Collapse -
Cnet anomalous
Aug 13, 2016 5:11AM PDT

I consistently test 30+ Mbps on other sites Cnet gives me 6 Mbps. Running Win 10. What's up with that?

- Collapse -
Cnet Test is Wrong
Aug 13, 2016 5:29AM PDT

I agree. Seems that Cnet needs to correct something in their test so that it will work with Windows 10.

- Collapse -
What tests can you believe?
Aug 12, 2016 5:45PM PDT

I have xfinity x1 which includes a speed consistently higher than the 120mbps they promise on okla's speedtest.net. Just now it is 177.29 down and 12.16 up. Right below the test results ookla says there is a faster connection available. I feel so inadequate.

- Collapse -
Rigged by the ISP
Aug 12, 2016 6:02PM PDT

Internet speed tests from the ISP are rigged by the ISP using their own server nearest to you. Better to use speedtest.com or ookla.com to get a result closer to what you actually experience when doing real tasks on the net.

- Collapse -
hold off on speedtest.com
Aug 12, 2016 6:20PM PDT

whoa there ben. big mistake if you type speedtest.com when you should enter speedtest.net. speedtest.com is bogus and rife with misdirection and who knows what else they spew.

- Collapse -
Least Resistance
Aug 12, 2016 6:15PM PDT

Many users have been critical of their ISP results differing so much from third party applications. Of course the ISP results will be better.

Of course the closest server will be used; microprocessors do that.

I have fiber right to my D-Mark yet I've never seen anything better than 37D. Totally acceptable. Do I always get a 37D? No, but it's not because my ISP didn't deliver on their promise.

- Collapse -
Absolutely
Aug 12, 2016 6:32PM PDT

Absolutely! Some ISP's apparently give you a high upload and download speeds for the first few Mbps Kb or whatever. Then throttle you down. I ran into this quite often when I had a slower service. I suspect that they still do it.

Spectrum ran a lot of line in our area and upgraded our band width to 100 Mbps. It appears that my download speeds are significantly improved.

After reading this post I ran a few tests. Speedtest.net connected to the closest major city and reported 130 Mbps down load and 5 upload. Cnet's Internet speed test reported 10.6 Mbps with no indication of this being upload, download or both. Who knows what my speed really is?

In general cable blows DSL out of the water. If you have the choice I strongly recommend cable. ( When ATT brags about their "high speed" don't believe it.)

- Collapse -
Lots of Slowdowns
Aug 12, 2016 7:28PM PDT

My day job and personal obsession is network and application performance for a very large global corporation. Key factors that affect your speed test performance:

1. Latency, or PING time. The higher the PING time between you and the test machine, the more important TCP tuning becomes. Speedtest.net does a good job configuring their test machines. If you see slower scores for higher PING times, it's time to improve your TCP tuning. See below.

2. Your broadband provider's peering arrangements. There are many carriers on the Internet, they pass off traffic at peering points, which are like highway intersections. Doesn't matter how many well-paved lanes both highways have if the cloverleaf intersection (peering point) is a rutted mud road. If traffic flow is uneven - more traffic flowing into Verizon's network from that other carrier, for example - then the peering point can become a p*ssing contest, where traffic is throttled while payments are debated. Found interesting results some years back, when Verizon wanted to punish customers for streaming off the reservation. There were Speedtest.net test points that gave good readings weekday mornings, but lousy results when the peering point was saturated that same evening. Try your test on a weekday morning to see if this is an issue with your carrier.

3. Your broadband carrier's network between that peering point and your house. Hotels and airlines overbook - sell 105 seats on a flight with 100 seats - since there will always be a few no-shows or last minute cancellations. Similarly, broadband carriers oversubscribe. They may have only 1 Gbit/sec capacity to your neighborhood, while total subscriber download capacity is far higher. They figure that very few people will use high bandwidth at the same time. However that link can get saturated during the evening. Continuing streaming growth, at ever-higher resolution, aggravates congestion.

4. Your home network. Most accurate reading is via Ethernet cable between a LAN port on your broadband router and your machine. Once you have that number, disconnect the Ethernet and test wireless. Then move further from the wireless access point and test again. If all readings are the same, you've got a fabulous home network. Otherwise, understand you may have "below rated" speed when you use wireless from your easy chair.

5. TCP tuning. Even Windows 10 has sub-optimal TCP tuning as installed. Macintosh is better. Only recently has iPhone TCP tuning allowed real high download speed. TCP auto-tuning is nice, but... your system has to be able to expand buffersize to 10 MBytes or larger. Bandwidth-Delay product isn't even table stakes, there's a lot more latency than just PING. Plus a few other knobs.

6. Oh, and not all speed test sites are created equal. Too-modest TCP tuning on some speed test machines means that remote users will never get good numbers. The speed test machine may have inadequate bandwidth, too-small memory, insufficient CPU power or use congested peering points. For example, CNET used one machine to cover the US, guaranteeing a lot of high PING time tests. Their machine wasn't tuned well enough to compensate for that high latency.

As another post noted, your broadband vendor's on-network machine will give best-possible results. Paying staff to handle bandwidth complaints costs more than giving the test machine(s) endless bandwidth, robust configuration and "turn it up to 11!" TCP tuning.

Sorry to run on, but there's a lot of slow between your keyboard and an Internet bandwidth test site.

- Collapse -
Yep They Are Rigged
Aug 12, 2016 8:10PM PDT

You bet they are rigged! That is to say, that the speeds boasted are those obtained under optimal conditions, especially during selected times when the number of users is low. Try running tests from independent web sites such as DSL Reports, etc, and you will see that the speeds vary all over the place. Low speeds are easily seen during periods of high numbers of users, and when businesses open in the various time zones. 'Too many users on too few routers' is the general cause of slow speeds.

Plus there are factors not related to the ISP's network but to the website computers which are limited to how much bandwidth they will allow (or are just over loaded with user demand). Witness Apple's servers when a major release of OS X comes out.

Like all advertising in the US, take everything with a large grain of salt!