Wednesday, June 26, 2013

The ISP Speed Claim Dilemma

ISPs face marketing issues no different than other providers of goods and services, namely that consumers generally have some expectations about what features and what prices constitute a reasonable offer.

That means every provider wants to appear to have an edge of some sort, and at a minimum, to supply the baseline of features consumers expect.

ISPs have one additional problem, namely that their product is essentially intangible. As with any other intangible product or service, a consumer cannot fully evaluate product claims until after the product is purchased.

But it still is reasonable to argue that most consumers considering the purchase of an Internet access service will evaluate the advertised speed and the advertised price. That means there will always be pressure to advertise the highest possible speeds.

But consumer protection agencies and regulators do not tend to like exaggerated claims. That is why more attention now is focused on how closely ISPs are able to deliver on speed claims. And there might be more work to do on that score in Europe than in the United States.

Actual European Internet access speeds are about 74 percent of advertised speeds during peak hours, a new study sponsored by the European Commission has found. Recent studies by the U.S. Federal Communications Commission have found that U.S. ISPs deliver actual speeds about 96 percent of advertised speeds.

The average download speed across all measured countries was 19.47 Mbps during peak
hours, and this increased slightly to 20.12 Mbps when all hours were considered.

Performance varied by access network technology.

Digital subscriber line services achieved 63 percent of the headline download speed, while cable services managed to achieve 91 percent of headline speeds. Fiber to home or VDSL services delivered 84 percent of headline speeds.

Fiber to home services achieved the fastest speeds in absolute terms, at 41 Mbps. Cable
services achieved 33 Mbps, whilst DSL services delivered 7 Mbps, on average.

In the September 2012 testing period, U.S. ISPs on average delivered 97 percent of advertised download speeds during peak periods, statistically equivalent to the last report, which found that the studied ISPs were able to deliver 96 percent of advertised speeds during peak hours of use, the FCC has reported. Those results were in line with testing conducted in 2011 as well.

On average, during peak periods DSL-based services delivered download speeds that were 85 percent of advertised speeds, cable-based services delivered 99 percent of advertised speeds, fiber-to-the-home services delivered 115 percent of advertised speeds, and satellite delivered 137 percent of advertised speeds, the FCC says.

This compares to July 2012 results showing largely the same performance levels: 84 percent for DSL, 99 percent for cable, and 117 percent for fiber. These results suggest that many ISPs are meeting established engineering goals for their respective technologies.

It isn’t immediately clear why DSL networks in the U.S. market were able to deliver real-world speeds more nearly matching advertised speeds, compared to European DSL networks.

But a reasonable guess is that the gap is explainable almost entirely by ISP marketing claims.
The EC study says two countries did not achieve 50 percent of advertised speed. Those two countries primarily use DSL networks, but more importantly “advertised their services
using only a handful of very high headline speeds.”

Hungarian ISPs delivered actual speeds that were 94 percent of advertised.  DSL services also achieved over 90 percent of advertised speeds. So it appears the difference is the marketing of service, not something inherent in the networks.

Since DSL performance is directly related to loop length, experienced speeds for consumers closer to the central office will increasingly diverge from speeds experienced by consumers further away from the central office.

It makes quite a difference whether the typical speed at 1,000 meters is used as the reference, compared to 5,000 meters.

But loop length is not the only consideration for DSL or other providers. Contention ratios and the degree of sharing also will affect performance. ISPs simply need to market services that reflect all the known limits, if they want to deliver on promised speeds.

But that is the dilemma. A more-realistic set of claims might mean forfeiting an advantage to other providers.


No comments:

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...