Mystery shoppers commissioned by Ofcom have found that 15 percent of the time, "potential customers" were not given an estimate of their access line speed, and 42 percent were only given one after prompting the sales agent near the end of the sales process.
The accuracy of the information proved to vary. In some cases, users were given double the line speed of another provider for the same line and technology, and sometimes received different answers over the phone when compared with the website of the same service provider. The majority of line speeds also did not match (within 1Mbps) the speeds given by the BT Wholesale line checker.
There is no question but that "best effort" broadband services are difficult to accurately predict or describe. It is true that users will sometimes experience bursts that correspond with the advertised "up to" speed. Most of the time, actual experienced sustained rates are lower, because of contention ratios and actual end user volume.
So Ofcom proposes that ISP's provide speed estimates based on line length, line capacitance and line attenuation, all measures that will provide a better approximation of typical download speeds.
Ofcom also wants to ensure that shoppers are given this information early in the sale process, particularly before payment information or a request for service is made.
Ofcom also seek to ensure that factors that affect broadband speed are explained. Specifically, Ofcom wants to ensure that shoppers are told how network capacity, congestion on the Internet and traffic management policies could affect performance. Consumers also should be told that actual throughput speeds will be lower than advertised or theoretical line speeds.
U.K. consumers already have the right to be moved to a cheaper, lower speed option when the plan they bought does not measure up. In cases where their is but one tier of service, Ofcom wants to allow consumers to leave their contracts without penalty.
Ofcom apparently will try to get such changes made voluntary. If changes aren't agreed to, or implemented, a regulatory review may occur, which could lead to formal regulation.
Such policies are not unreasonable consumer protection efforts. The problem is that formal guarantees are next to impossible so long as connections operate on a "best effort" basis.
Even on a quality-assured connection, which would have to be based on packet prioritization policies, throughput will vary throughout the day, based on overall contention for network resources, though far less than is the case on a "best effort" connection.
To closely match expected routine performance with an advertised top speed will require packet prioritization.
source
