|
The FCC voted recently to implement consumer broadband labels. This was required by section 60504 of the Infrastructure Investment and Jobs Act. The new rules will become effective after the Office of Management, and Budget approves the new rules and after the final notice is published in the federal register. ISPs will then generally have six months to implement the labels.
The labels look a lot like the nutrition labels that accompany food. The label will include basic information like a customer’s service plan name, the monthly price for standalone broadband, any special pricing that is in place currently and when that special pricing expires, and a description of other monthly and one-time fees. The labels also must disclose the typical broadband speeds and latency.
It’s going to be interesting next summer to see how ISPs react to the label requirement. The pricing information alone must be giving shivers to the marketing folks at the biggest ISPs. This will make them list the price of standalone broadband to every customer and compare that to what the customer is currently paying. It’s been incredibly easy for consumers to subscribe to broadband in the past and never know the list price of what they are buying.
The requirement that I think will be the most controversial is the requirement to disclose the typical broadband speed and latency. I can’t wait until next year to see big ISPs implement this requirement. In the many surveys we have done, most consumers tell us that they have no idea of the speeds they are supposed to get—and that most of their monthly broadband bills don’t mention the speed.
Some ISPs will have a real dilemma with the speed disclosure.
We’ll have to wait and see, but my guess is that ISPs will report the same speeds to customers that are reported to the FCC. But an ISP exaggerating FCC speeds should be ready for an onslaught of customer complaints from customers who know the speeds on the label are not right. I think this is part of the reason why these labels were mandated—to force ISPs to come closer to telling customers the truth. But there are going to be some contentious years coming for ISPs that claim imaginary speeds on the broadband label or to the FCC.
The FCC is not quite done with the labels. The FCC issued a Future Notice of Proposed Rulemaking to solicit input on a few issues. One is how to include bundling on the labels. The surveys my firm does are still showing more than 50% of customers on bundles in urban areas, and bundling allows ISPs to hide the pricing for any component of the bundle.
The FNPRM also asks if there is a better way to disclose speeds, such as average speeds instead of typical speeds. Finally, the FCC is asking if ISPs should disclose network management policies that might harm consumers, such as blocking, throttling, or paid prioritization for some customers.
We won’t see the broadband labels in practice until at least next summer—but I’m expecting an uproar after folks see what ISPs say about prices and speeds.
Sponsored byCSC
Sponsored byVerisign
Sponsored byIPv4.Global
Sponsored byDNIB.com
Sponsored byRadix
Sponsored byVerisign
Sponsored byWhoisXML API
The speed value is partially meaningless and is certainly vague.
For instance, if the MTU on access path is less than 1500 then a lot of packets might get fragmented.
And when one talks of bits or bytes per second, does this include Ethernet frame headers, Ethernet CRC, IP headers, TCP headers, UDP headers or what? This especially affects traffic that uses small packets, such as VoIP where a significant percentage of the bits being moved might be in those header fields that are, or are not, counted in the “bandwidth” value.
And what about asymmetry of up and down rates? (And, as you mention, the change in rates due to time of day or other users on the physical packet distribution tree.)
Also of interest is what happens to broadcast or multicast packets - do others on a cable plant see those?
“Typical latency” is also excessively vague. Latency up or down? And latency alone is not all that useful without a measure of the shape (burst characteristics) and statistics of variation of that latency.
Also needed are things like buffering - does the path have too few or too many (bufferbloat) buffers, and if two few what is the packet drop policy?
This is the kind of inadequate disclosure that happens when the designers of that disclosure only view the world via things like Ookla speedtest.