By Phil Goldstein
Wireless carriers like to brag about their LTE networks in a lot of different ways, including their coverage, reliability and speeds. And while the average speed for downlink and uplink are probably the most important metrics to judge carriers' performance on--since that is what a consumer can expect to experience on a regular basis--maximum speeds are also worth noting.
Maximum actual speeds for downlink and uplink performance show the limits of the network. They demonstrate the engineering that carriers have engaged in to really let customers put the pedal to the metal. Such maximum speeds are also highly dependent on how much spectrum carriers are devoting to LTE and can provide a window into how carriers are using their spectrum resources. They are also dependent on a host of factors including congestion on the network, the user's proximity to the cell site, and other elements.
Carriers sometimes tout the maximum speeds their networks can afford subscribers. T-Mobile US (NYSE:TMUS), for example, noted in its third-quarter earnings materials that customers in markets where it has deployed at least 15x15 MHz of spectrum for LTE are seeing real-world peak downlink speeds of up to 145 Mbps. But how are the carriers really doing in pushing the envelope on speeds?
Thanks to an exclusive partnership between FierceWireless and network testing firm RootMetrics, readers will be able to see the maximum recorded downlink and uplink speeds the firm observed for the four Tier 1 carriers across eight different regions in the U.S. The data also lets readers see which regions of the country in general let wireless customers access the highest peak downlink and uplink speeds. RootMetrics conducted extensive testing across the country both in the first half of 2014 and in the third quarter to gauge how the four Tier 1 carriers fared in terms of maximum downlink and uplink speeds.
Those figures could indicate several things. Verizon has been busy deploying its AWS spectrum as part of its "XLTE" service for additional capacity, which could provide higher peak speeds. In September, Verizon said it had covered 400 of its 500 LTE markets with AWS deployments, in some markets giving the company an extra 20x20 MHz of capacity. That deployment likely had a large effect on Verizon's maximum speed numbers.
Sprint, on the other hand, has been building outs its tri-band Spark LTE network at a relatively slow pace. Sprint says the network can deliver peak speeds of up to 60 Mbps. However, in the third quarter Spark was available in just 29 markets, limiting its reach (Sprint has since expanded its Spark service to 46 total markets and at least 92 million covered POPs, with the goal of hitting 100 million POPs by year-end). In the third quarter the highest maximum observed downlink speed RootMetrics observed on Sprint's network was 30.35 Mbps, in the Rocky Mountain region. Part of the reason for that could be that Sprint's larger LTE network, on its 1900 MHz spectrum, is being deployed using a 5x5 MHz block of spectrum, which limits downlink speeds.
All of the carriers saw a general improvement in their maximum speed performances for uplink and downlink speeds from the first half to the third quarter, according to RootMetrics, though that could be a quirk of the data, as explained below.
In addition to studying the individual carriers' performance, RootMetrics also drew general conclusions about the speeds observed in each region. For example, New England was the "fastest" region overall, offering a 51.75 Mbps weighted maximum downlink speed and 22.5 Mbps weighted maximum uplink speed. The Great Lakes region came in second (48.8 Mbps weighted maximum downlink and 19.21 Mbps uplink weighted maximum uplink). The Plains came in third, the Southeast came in fourth, while the Southwest came in last, with a weighted maximum downlink speed of 38.47 Mbps.
There are some caveats and information readers should know about the data (see below for more details on RootMetrics' methodology). RootMetrics is comparing maximum observed downlink and uplink speeds in two time periods--the first half of 2014 and the third quarter of 2014. However, not all 125 U.S. metro areas RootMetrics surveys are included in the third-quarter data; the third-quarter numbers include 52 U.S. metro areas, while the first-half figures include all 125 metro areas. That is simply because in the third quarter RootMetrics did not collect survey data for all of the 125 metro areas. The firm says it will be able to include more when it collects the data again in the fourth quarter.
In terms of RootMetrics' methodology for weighting and comparing regions, the firm notes that a regional weighted maximum upload and download speed was found by averaging a carrier's speeds, weighted by the proportion of overall subscribers for that carrier. The regional comparisons were then made using this weighted number.
Here are the maximum sustained upload and download carrier speeds, broken down by U.S. region, as reported on by RootMetrics in the third quarter. Note: These are graphics created by FierceWireless. RootMetrics has not drawn the conclusions illustrated below, and does not endorse one carrier over another. Your results may vary.
And here is the full report from RootMetrics:
|Region||Carrier||1st Half||3rd Q||Diff||1st Half||3rd Q||Diff|
For a state-by state breakdown, this is how RootMetrics defines each region:
- New England: Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island and Vermont
- Mid-Atlantic: Delaware, District of Columbia, Maryland, New Jersey, New York, and Pennsylvania
- Great Lakes: Illinois, Indiana, Michigan, Ohio and Wisconsin
- Plains: Iowa, Kansas, Minnesota, Missouri, Nebraska, North Dakota and South Dakota
- Southeast: Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, Virginia and West Virginia
- Southwest: Arizona, New Mexico, Oklahoma and Texas
- Rocky Mountain: Colorado, Idaho, Montana, Utah and Wyoming
- West: Alaska, California, Hawaii, Nevada, Oregon, and Washington
Analysis of maximum sustained data speeds was conducted to estimate the highest speeds likely to be observed and maintained for a period of time. Maximum sustained speed calculations remove outlier results that are not likely to represent expected consumer experience. Exclusions occur based on the following two criteria:
1. Results more than 5 times above the inter-quartile range (IQR) of the median were excluded for each carrier market.
2. For each region, the top 5% of all remaining speed tests were analyzed to determine if surrounding speed tests in a 30-minute interval were within 50% of the candidate speed test. If they were not, those anomalous test results were excluded from consideration.
The top remaining speed test was considered the maximum sustainable speed for a region.
Article updated Dec. 14 with additional clarification from RootMetrics.