Broadband Internet service is still painfully expensive in the United States and to make matters worse, a new study shows that most American households aren’t getting the fast data speeds they are paying for.
The Wall Street Journal on Monday issued a report that included a great study from Ookla, operator of the popular data speed testing website Speedtest.net. The data included the results of tens of millions of speed tests as well as a survey of 646,404 Speedtest.net users taken over the past 12 months.
The purpose of the study was to determine how close Internet service providers in the U.S. come to matching the data speeds their customers are paying for. In other words, are you really getting download speeds of 30Mbps when you pay for that 30Mbps broadband package?
More often than not, the answer to that question is no.
Ookla’s data shows that ISPs covering the majority of the country provide customers with data speeds that fall short of the broadband speeds they advertise. Only six ISPs — Midcontinent Communications, Earthlink, Optimum Online, WideOpenWest, Verizon FiOS and Charter Communications — offered data speeds that were faster than advertised on average.
Meanwhile, subscribers to broadband service supplied by 20 top ISPs in the U.S. saw data speeds that were slower than advertised.
Some highlights among top ISPs:
Optimum Online customers saw speeds that were 3% faster than advertised, on average, and Verizon FiOS subscribers averaged about 1% over advertised speeds. Average data speeds seen by Cox customers were exactly in line with advertised speeds.
On the other side of the fence, data speeds for Time Warner Cable customers were 1% slower than advertised, and Comcast speeds averaged 2% slower than advertised. AT&T U-verse data speeds were 8% slower than they should be according to Ookla’s data, and Clear Wireless was the worst of all the ISPs tested, with download speeds that were 41% slower than advertised.
The one giant caveat in this study is obviously that it is completely unscientific. The speed tests used to arrive at these conclusions were not controlled at all. There is no way of knowing how many users were connected to their routers via ethernet and how many were connected wirelessly or through third-party modems, which might impact observed data speeds.