Charter defense that everyone’s broadband is slow just won’t fly

The big mistake that Charter Communications has made in its suit from the New York Attorney General, which claims it misled consumers with advertised broadband speeds, was to fight it at all.

But the company was Time Warner Cable then, and it had already dug in its heels and claimed that black was white and denied everything. What it might have done is gone for mitigation as the new parent arrived, agreeing to a settlement and also promising to work with the Attorney General to replace the 900,000 or so modems at the heart of this disagreement. It may not be too late even now.

The New York case dates back to 2012, but really materialized in 2016, when some 640,000 consumers were reported as receiving slower speeds than promised, and said they even had trouble accessing content from Facebook, Google, YouTube, Netflix and various gaming platforms. This was put down to some 900,000 older generation modems which were still in situ and given that Charter fully intends to replace these anyway and mentioned this very same fact at its Q4 results briefing, as it plans a 1 Gbps broadband service. This translates to moving from DOCSIS 3.0 to DOCSIS 3.1.

The problem here is that US cable universally talks about the standard, and not about the CPE that is delivering it. So throughout the US in DOCSIS 3.0 environments, typically the cable companies only bought 300 Mbps home gateways. DOCSIS 3.0 spectrum begins at 54MHz and can go all the way up to 1002 MHz. Its theoretical limit is 1.2 Gbps real world downstream with upstream capacity capped at 216 Mbps. But no cable firm in the US delivers this but promise it all the time. To do that it would have to use all the spectrum and that would mean the price of a CPE which could handle it would go through the roof. The spectrum is organized into groups of 32 channels of 6 MHz each. But more than one of these groups can be activated, depending on the capacity of the CPE. It tends to be a 200 to 300 Mbps service in the real world.

DOCSIS 3.1 is a complete step change which allows for really modular channels of around 50Khz each, and you can use as many as you want because it uses OFDM instead of QAM – multiple signals modulated separately instead of one big signal modulated aggressively. DOCSIS 3.1 has a theoretical upper limit of 10 Gbps downstream and two Gbps uplink. This can be organized into 1,2 3, or 4 chunks of 3840 sub-carriers, with about 70% of the performance in real data delivered. This means most DOCSIS 3.1 CPE will support just one of these sets of sub carriers and throughput will be about 1.8 Gbps down, and 350 Mbps back. But all cable firms and the technology companies behind DOCSIS 3.1 keep saying it is 10 Gbps, when only about 70% of that performance is available as data and often only 25% of that maxes out most CPEs.

Comcast has just begun offering a DOCSIS 3.1 service with one Gbps download and 35 Mbps upload. Quite appropriately this is the advertised speed, not 10 Gbps. The other numbers are really about seeing off the Telcos who are just moving from 50 Mbps to 100 Mbps, and then onwards to 1 Gbps, by embracing technologies like G.Fast, VDSL2 35b and fixed wireless 5G.

As AT&T and Verizon get closer to matching the service of cable firms, there will be more mis-selling of broadband speeds, not less.

The key to the case, and how much Charter will end up with as a New York State fine if it loses it, will come down to how much slower the Time Warner Cable devices were compared to other cable firms. New York state says – “a lot”.

The big problem is that the 200 Mbps service that it was offering was reported at offering some 30% as its top speed , around 60 Mbps. This is ironic in that in his speech to investors last week, CEO Tom Rutledge talked about upgrading the old 60 Mbps modems to newer 200 Mbps devices. So if the NY Attorney General is right, some people would have upgraded to the new service expecting 200 Mbps and would have got the same speed as before. That’s got to be disappointing.

The judge has been asked to throw out the case previously because it should have been a federal case, not a state case, and so should be outside the scope of the NY Attorney General, and that idea was rejected and what happened this week is that Charter failed to convince the court that its claims and broadband performance were the same as everyone else’s.

It doesn’t make it any less worrying that it was the old weak management of Time Warner Cable, that “ripped off” New York customers, and that the new management plans to fix things, since the first thing the new management did was move people onto a unified product that cost more, exacerbating the problem. This whole problem dates back to 2012, and had Charter come clean and worked on fixing the problem, a fine may have been avoided.

Now a court will later contemplate how much money Charter should return to all of these 640,000 clients in rentals of devices and service payments going back to 2012. We suspect it could be as much as half the money paid out if Charter loses.

If the issue of Time Warner Cable throttling Netflix is further proved, punitive damages would make this case very expensive for Charter and it will take a long look at its agreements to purchase the company and see what provisions are contained in them to claw back some of the $48.5 billion it paid for the business.

The claim is that clients had trouble accessing Facebook, Google, YouTube and Netflix. A broadband line that has trouble with all of these services is now worth the name “broadband.” But one other area that may prove really contentious in all of this may be the quality of the WiFi chip used inside the older modems. This could have been 802.11n or worse, 802.11g which ran at a measly 54 Mbps in congested 2.4 Ghz spectrum. If that’s the case, the performance is likely to be way lower than 30% of the advertised rate.

This shows how important recent contracts (we cover one in the issue this week with Deutsche Telekom) are for super high speed WiFi, which also has multiple Access Points, and dynamic band and client steering, as well as insights into how it’s performing that can be seen from the help desk on cloud apps.

Charter’s last quarter shows that it spent almost $700 million more on capex than previous chasing the problem and it may have to continue to spend at that level for the coming year, including improvements to its basic in-home WiFi, in order to put it past this problem, and begin to build its reputation among TWC customers, once again.