Should Grant-funded Networks be Open-Access?

There was an interesting political effort in the Washington legislature recently to expand the use of open-access networks. There was language included in the Substitute House Bill 1147 that would require that any network funded from BEAD grants must become open-access and available to other ISPs.

Open-access has been a topic in Washington for many years. There was a long-time prohibition against Public Utility Districts (PUDs) from offering retail broadband. These county-wide government-owned utilities wanted to bring better broadband and settled by building open-access networks. Over the last few decades, a number of PUDs have launched open-access fiber networks, some of them with tens of thousands of fiber customers.

For those not familiar with open-access, it is a network where multiple ISPs can buy access to reach customers. This provides customers with the choice of using multiple ISP on the same network. All reports are that customers like the extra choices they get. Every broadband survey my firm has ever conducted has shown a huge public preference preference for choice.

The legislature finally relaxed the prohibition for PUDs last year, but in a bizarre fashion. The legislature passed two conflicting bills that allow PUDs to provide retail broadband services. Rather than choose between the two bills, the Governor signed both simultaneously (a pen in each hand) so that both bills went into effect. As might be imagined, this created as much confusion as clarity over the issue.

I doubt that anybody will be surprised that the biggest ISPs in the state vehemently opposed this legislation. The big cable companies have always immediately fought any suggestion that they allow other ISPs to use their networks. The big telcos were forced to sell unbundled copper loops starting with the Telecommunications Act of  1996, but that requirement continues to wane as the amount of copper keeps shrinking. The telcos started fighting against the unbundling rules as it was enacted, and over the years succeeded in greatly weakening the ability of outsiders to use their copper.

I don’t think anybody will be surprised to find out that the big ISPs in Washington succeeded in killing this idea. The big ISPs threatened to not pursue any grant funding if this proposal becomes law. Some even made veiled threats to stop investing in the state if this became law.

But it’s an interesting concept. The BEAD grant rules have a clear preference for open-access networks, and any carrier promising an open network will get extra points on a grant application. But the open-access preference is only a suggestion and not a requirement – something the big ISPs in Washington all pointed out.

Requiring open-access is not a far-fetched idea because open-access is required on all of the middle-mile networks that were announced this week as recipients of NTIA grants. But the whole point of the NTIA middle-mile networks is to build networks to places where backbone connections are unavailable or unaffordable. Requiring the grant recipient to sell affordable connections to everybody is a good use of federal grant dollars.

But this raises a much larger question. I know there are a lot of open-access proponents in the country who think that any network funded with government dollars ought to be made open-access to provide the most value to the taxpayers who are funding it. That is exactly what was suggested in Washington, but it didn’t take very long for the big ISPs to kill the idea.

Many industry folks want to take this idea even further. I don’t think I’ve seen a thread on this topic that doesn’t include somebody who thinks government should own all grant-funded fiber infrastructure, which should then be made available to all ISPs that want to use it. Obviously the BEAD grant rules weren’t written that way, and with the sway that big ISPs hold in D.C. it probably never will happen. But is something that Congress could do if they ever have the will to enact it. We’re starting to see cities who are adopting this idea, so we’re going to keep seeing new open-access networks coming to life. I have to think that the citizens in every city close to an open-access network is going to be asking why they can’t have the same thing.

The FCC to Look at Data Caps

FCC Chairwoman Jessica Rosenworcel asked the other Commissioners to join her in opening an investigation into broadband data caps. According to FCC rules, a majority of Commissioners must agree to open any official proceeding. For those not familiar with the data cap concept, it’s where an ISP bills extra for using more than a defined amount of broadband in a month.

Not all ISPs use data caps. The ISP that gets the worst press about data caps is Comcast, but it doesn’t bill data caps in all markets – seemingly only where it doesn’t have a lot of competition. Charter would love to bill data caps, but it has been prohibited from doing so because of an arrangement reached with the FCC when it got approval to buy Time Warner. That agreement just lapsed on May 18 of this year. We’ll have to wait to see if Charter will impose data caps – but it seems likely it will do so since the company asked permission from the FCC to impose data caps in 2021.

AT&T imposes Data caps on DSL and on some fiber connections. Astound broadband charges data caps in Washington, Oregon, and California. Cox has data caps that kick in after a user exceeds 1.25 terabytes per month. Mediacom imposes data caps on many of its plans. All of the products of the high-orbit satellite companies, HughesNet and Viasat, have severe data caps. So do cellular hot spot data plans.

It’s an interesting request by the Chairwoman. Under current FCC rules, the FCC has no authority to do anything about data caps. This authority went away when the previous FCC under Chairman Ajit Pai eliminated the regulation of broadband by killing Title II authority. Chairman Pai went even further and pushed remaining vestiges of any broadband regulation to the Federal Trade Commission.

This makes me wonder why Chairwoman Rosenworcel would try to open this docket. I can see several possibilities. First, this could just be done to show that the FCC cares about an unpopular ISP practice. It’s clear that the public hates data caps. I saw that the press that covers the FCC immediately flooded the news after this was announced. I would hope the Chairwoman would not be so callous as to investigate something for which the FCC is powerless to make any changes.

That leads to the second possibility that Chairwoman Rosenworcel believes that adding a fifth Commissioner will provide the votes needed to reinstate Title II authority or some updated version of it. Starting the investigation into data caps now might sync up well with renewed FCC regulatory authority and let the FCC make a popular change in the future to ban or modify data caps.

I’ve written several blogs over the years that make the argument that data caps are nothing more than a way for ISPs to extract extra payments from customers. There is zero justification from a cost perspective that residential customers that use more data than average cause any significant incremental cost for an ISP. ISPs buy wholesale broadband based on the busiest times of the usage in a month. Within the pile of broadband purchased to meet that peak need, it doesn’t matter how much broadband customers use as long as it doesn’t push up the busy hour for the month.

Additionally, the big ISPs that use data caps also engage in peering arrangements where they directly hand off broadband traffic to the largest web services like Google, Netflix, Facebook, and Microsoft. While there is a cost to create the peering points, once established, the amount of data sent through peering arrangements saves a huge amount of money compared to shipping this same traffic through the Internet.

It’s harder each year for affected homes to avoid data caps. Data caps accumulate both download and upload usage, and homes are increasingly using upload bandwidth that most folks don’t even realize. According to OpenVault, the average home in the U.S. now uses over 560 megabytes of data per month, an amount that keeps climbing. The household average as recently as 2017 was only 273 megabytes per month.

This will be an interesting process to watch. Chairwoman Rosenworcel has created a form for folks to describe their data caps stories. I’m sure that everybody who does so will be hoping that the FCC can help them – but that remains to be seen. It means getting a fifth Commissioner who is willing to reintroduce broadband regulation – something that is going to have a lot of opposition.

The Remaining RDOF Funds

The FCC originally budgeted $20.4 billion dollars for the RDOF subsidy program to be spent over ten years. The original RDOF reverse auction offered $16 billion in subsidies. But in a story that is now well known, some entities bid RDOF markets down to ridiculously low subsidy levels, and only $9.4 billion was claimed in the auction. $2.8 billion of this funding ended up in default, including some of the bidders who had driven the prices so low.

That means that only $6.4 billion of the original $20.4 billion has been allocated. The question I’m asking today is what the FCC will do with the remaining $14 billion.

It seems unlikely that there will ever be another RDOF-like reverse auction. RDOF was meant to bring broadband to areas that were unserved according to the FCC’s broadband maps at the time of the reverse auction – meaning areas where no ISP claimed broadband speeds of at least 25/3 Mbps. But since ISPs are able to claim marketing speeds under the FCC mapping rules instead of actual broadband speeds, many millions of unserved locations were left out of the RDOF process.

Since the RDOF auction, there have been many billions spent to bring broadband to unserved areas through ReConnect grants, local ARPA grants, state broadband grants, and several smaller grant programs. To understand how poor the original FCC RDOF maps were, even after these many grants, the latest FCC maps still show over 8 million unserved locations. Folks like me who look at the map at a granular level think there are even more areas that are still mistakenly claimed to have 25/3 Mbps broadband but that don’t in real life.

To be fair, the RDOF is doing some good things. A lot of electric coops, telephone coops, telephone companies, and independent fiber overbuilders are building networks using the RDOF subsidy as the basis for getting funding. Charter and a few other larger ISPs are building networks using the RDOF funding.

But the RDOF awards also left behind a lot of messes. First, it took too long to eliminate the default bidders. Areas claimed by these bidders were off-limits to other federal grants and most state grants – many of these areas would have fiber today had they not been in RDOF limbo.

The bigger problem is that the FCC made an absolute mess by awarding RDOF in what can be best called a checkerboard RDOF serving area. The following map is a good example of what this look like in a real county. In this particular county, the areas to the east have no people due to large parklands, but in the rural areas where people live, the RDOF awards covered some areas but not adjacent Census blocks. The Census blocks that were not awarded have the same lousy broadband options and were not included in the RDOF award due to the mapping problems discussed earlier.

This creates a real challenge for anybody now trying to get a BEAD or other grant to serve what is left. The areas left after RDOF don’t make a big coherent serving area, but a jumbled mess of remaining Census blocks. For somebody building a fiber network, these checkerboard areas are a nightmare because a builder must go through RDOF areas to reach the remaining areas. It’s one more factor that will drive up the cost of the BEAD grants in counties that got a lot of RDOF funding.

The FCC is dreadful at awarding grants and subsidies. The RDOF process was used so the FCC didn’t have to review traditional grants where ISPs proposed coherent grant serving areas. This is the same FCC that gave over $11 billion to the biggest telcos for CAF II to upgrade DSL to 10/1 Mbps.

Now that the states have broadband offices, the easiest way for the states to award the remaining RDOF billions would be to let state broadband offices do the heavy lifting. It would be one more tool for state broadband offices – that hopefully would not follow the complicated BEAD rules. The worst possible way to use the money would be for the FCC to take some easy path to shovel the money out the door again – please don’t give us RDOF II!

New Broadband Trends

The latest Broadband Insights Report is out from OpenVault providing statistics on average broadband usage at the end of the first quarter of 2023. In looking over the latest statistics I’m starting to see some interesting trends.

The average household used 560.5 gigabytes of broadband per month by the end of the quarter. That is the combination of 524.8.2 gigabytes of download and 35.7 gigabytes of upload. I also looked back over past years, and I think a new trend of broadband growth is emerging. Consider the following simple tables based upon the average household usage at the end of the first quarter since 2019.

Gigabytes
1Q 2019 273.5
1Q 2020 402.5 147%
1Q 2021 461.7  15%
1Q 2022 513.8  11%
1Q 2023 560.5  9%

2020 growth was crazy due to the pandemic, and that level of growth is likely never going to be seen again absent some other similar catastrophic event. Since then, growth has slowed a bit year after year. We’re settling into a pattern where the average household is using approximately 50 gigabytes more per month than the year before.

This is easy to understand. More and more things we are moving online. We’re storing more pictures and files each year. We’re watching more videos, and those videos are growing more data intensive as we migrate to 4K video. Most of the software we use is now in the cloud. The devices in our house are often connected to the cloud.

This is starting to feel like a new trend. We used to have a paradigm that broadband usage doubled every 3-4 years. Once we’ve moved most of our data lives to the cloud, there is no longer the likelihood of explosive growth in household usage. I think more and more homes are settling into a mature stage of having moved to the cloud. The only thing that can upset this would be some new widely used data function that uses a lot more data, or another event similar to the pandemic.

The other trend is that people have wholeheartedly decided that they want faster broadband speeds. There are a lot of folks in the industry who will argue vehemently that households don’t need more than 25 Mbps – but it doesn’t matter what they think. Huge numbers of families believe they should have faster speeds and are upgrading. Consider the following table that compares the percentage of subscriptions to various speeds from the first quarter of 2021 and 2022.

 Subscribers

1Q 2022 1Q 2023
Under 50 Mbps 7.6% 4.7%
50 – 99 Mbps 6.3% 4.8%
100 – 199 Mbps 17.0% 9.3%
200 – 499 Mbps 49.7% 41.1%
500 – 999 Mbps 6.1% 22.0%
1 Gbps+ 13.4% 18.1%
200 Mbps + 69.2% 81.2%

The percentage of homes that are subscribed to 200 Mbps or faster has skyrocketed in one year from 69% of homes to 81% of homes. Some of this increase comes from ISPs arbitrarily increasing speeds for customers, but a lot of the growth comes from people deciding to upgrade. This is clearly now a major trend.

The day of talking about 100 Mbps being an acceptable broadband speed is now behind us when 81% of the homes in the country are subscribed to speeds at least double that. As usual, the politicians who wrote the rules for the BEAD and other federal grants are far behind the real-life curve. Grants that allow somebody to build a network that can deliver only 100 Mbps are investing in obsolete technology. By the time those grant networks are constructed, any new networks that deliver only 100 Mbps will be years behind the rest of the broadband in the country. Let’s hope that broadband offices pay attention to this trend and require technologies that are forward-looking rather than buried in the past before they are even constructed.

The Benefits of Thinner Fiber

Fiber manufacturers are always trying to make it easier to deploy fiber. One of the most interesting trends is the increasing migration from 250-micron fiber to 200-micron fiber. For those not familiar with the metric system, a micron is one-thousands of a millimeter. A 250-micron fiber has a diameter of 0.25 millimeters, while a 200-micron fiber has a diameter of 0.2 millimeters.

That may not sound like a big difference, but when each fiber is thinner, the overall size of a fiber bundle is smaller. A larger fiber bundle of 200-micron fiber can be 20-30% smaller than the equivalent bundle of 250-micron fiber.

200-micron fiber has been around for five years, but it’s growing rapidly in popularity. It’s estimated that as much as 10% of all fiber sold is now 200-microns.

Interestingly, the core glass in the two types of fiber is identical, with a 9-micron core and a 125-micron surrounding glass. The difference in the overall fiber size is due to different coatings. The identical core means that it’s possible and easy to fuse a 200-micron fiber with a 250-micron fiber – allowing a fiber builder to mix the two kinds of fiber in a network.

There are big benefits of using smaller fiber. The smaller size means smaller fiber bundles that are far easier to use during the construction process. There is the added benefit of getting a lot more of the thinner fiber on a reel, meaning fewer changes of reels.

One of the more interesting benefits is to use a bendable version of the thinner fiber (Bending Inflexible Fiber). Bendable fiber is more tolerant of bending during the construction process, leading to less damage and stress on the fiber. Thinner fibers make it easier to install indoor fiber in places with a lot of 90-degree bends.

Perhaps the ultimate benefit is that smaller diameter fiber makes it easier to use microduct conduits. If using 200-micron fiber, it’s possible to fit a 432 fiber into a 10-millimeter microduct or an 864 fiber bundle into a 14-millimeter microduct. This opens up the possibility of installing huge numbers of fiber along a street in those situations where it’s needed.

Who’s the Top ISP?

Americans have an insatiable desire to rate things. We want to go to the best restaurant in town and drive the most highly-rated car. PCMag ranks the biggest ISPs every year.  It’s an interesting exercise in that each year it points out a few noteworthy ISPs. The magazine asks readers to rate ISPs on a number of factors, including broadband speed, satisfaction with the connection process, and the interaction with customer service.

This year’s winner and highest-rated ISP is NextLight, the municipal fiber utility in the city of Longmont, Colorado. The ISP had the highest rating ever for PCMag and got an overall rating of 9.9 out of 10. The highest-rated national fiber providers were AT&T Fiber and Verizon FiOS, both with a rating of 8.2. Google Fiber was also highly-rated but doesn’t have a big enough footprint to be considered a national provider.

The highest rating for a cable company was Astound Broadband, with a rating of 7.8. Starlink was rated as the best satellite provider with a rating of 8.2. The final category was 5G FWAS wireless, and Verizon won in this category with a ranking of 7.9.

This is an interesting survey, but it is not close to being a scientific sample. The only voters are PCMag customers who also subscribe to the company’s What’s New Now newsletter. An ISP had to receive at least 50 customer responses to be considered. PCMag runs numerous polls throughout the year to let customers rate a range of technical products and services.

There are a few takeaways from the poll responses. The survey rated a local fiber provider the best with a rating of 9.9, while the best cable company in the country got a 7.8 rating. I can’t help but wonder how many other small fiber ISPs would get a high rating if they could get fifty customers to respond to the survey. It’s not hard to imagine that many other municipal providers, cooperatives, independent telephone companies, and small fiber overbuilders would have rated higher than AT&T and Verizon fiber, and higher than all of the cable companies.

The ratings for Starlink and Verizon FWA fixed wireless are interesting because these companies are bringing broadband to many places where all other broadband alternatives are inadequate. I talk to rural folks all of the time who have converted to FWA wireless, and they are thrilled to finally have a broadband solution that works. When you’ve been on rural DSL, high-orbit satellite, or a cellular hotspot, a new alternative is a welcome new option. But I have to wonder how well Starlink and FWA cellular wireless will do after federal grant funding results in fiber networks being built in many of the rural markets in the country? How many people will ditch Starlink and change to a fiber connection that’s both faster and a lot less expensive?

I have my own definition of the best ISPs. The best ISPs are those with almost no churn. This means customers don’t leave these ISPs to change to another ISP. Every ISP has churn from customers that move, die, or have a financial downside. But the best ISPs rarely lose customers to their competition. I’m sure this is true for Longmont’s NextLight as it is for hundreds of other small fiber ISPs across the country. Getting this accolade is great for Longmont. The many other small fiber ISPs don’t need a national magazine to tell them what a great job they are doing. Their customers show them that every day.

Frontier Plans to Kill Copper

CEO Nick Jeffery of Frontier said at a recent investor conference that the company believes it will be out of the copper business within five years. The company is facing the same dilemma as the other big copper owners like AT&T, Lumen, Verizon, and Windstream.

The company has an interesting path ahead to get rid of copper. At the end of the first quarter of this year, Frontier still had 9.9 million copper passings compared to 5.5 million fiber passings. But that oddly doesn’t translate into a greater number of copper customers, with 1.6 million fiber customers compared to just under 1 million copper customers.

This demonstrates the extent to which Frontier has lost DSL customers over the last decade. For much of the last decade, Frontier was the biggest percentage loser of broadband customers almost every year and quarter. At the end of 2017, the company had over 3.9 million broadband customers compared to 2.6 million now. Frontier shed some customers in the asset sale in the Northwest to Ziply, but most of the customer losses are from DSL customers fleeing to some other technology.

The company announced a goal last year to reach 10 million fiber passings by the end of 2025, so it will continue to overbuild copper areas with fiber. That should cover about half of the remaining copper passings. When Jeffery talks about getting out of the copper business, he’s talking about eventually walking away from the remaining 4 – 5 million copper passings.

But that may not turn out to be as drastic as that sounds. Many of Frontier’s copper passings will be overbuilt with fiber or some other technology as a result of the BEAD and other federal and state grant programs. Frontier will likely be participating in many of these grants to reach it’s goal of 10 million total fiber passings.

The other technology that has to be putting a big dent in DSL is FWA wireless from cellular companies. The pricing is similar to DSL pricing, but the speeds are faster. The cellular companies are marketing to the same demographic that has stayed with DSL even where cable broadband is available – households for whom price is more important than broadband speed. Verizon and T-Mobile have sprung from nowhere to gain 4.1 million FWA customers nationwide over the last year or so – and a lot of those customers must be switching from DSL.

My guess is that Frontier won’t have to cut many DSL customers dead in five years. Between their own fiber expansion, the many grant programs, and FWA cellular wireless, it seems likely that most of Frontier’s remaining DSL customers will be off copper by then. But there will inevitably be some unlucky remaining customers who will get the notice that their copper will be going dead with no offered replacement. My best guess is that Frontier’s exit from copper might be relatively easier than if the company tried to kill copper today, as is being done by AT&T and Verizon. Those two telcos are taking a lot of grief when they discontinue active copper customers. But in five years it’s likely that very few people will even be interested in the remaining Frontier copper. My guess is that market forces will get the company out of the copper business without too much pain.

Our Fixation on 25/3 Mbps

Mike Conlow wrote a recent blog on Substack that discusses how cellular companies are reporting large numbers of passings on the FCC maps as having the capability to receive exactly 25/3 Mbps or 100/20 Mbps.

Mike uses the example of North Carolina to highlight the issue. In the previous FCC map, UScellular claimed 1.13 million locations in the state that could receive speeds of 25/3 Mbps for fixed cellular broadband. In the latest map, UScellular dropped this number to 224,000 locations. T-Mobile made a similar claim, but only dropped claims of 25/3 Mbps capability from 1.06 million to 1.04 million.

That isn’t a very fast broadband speed, so why does this make any difference? It turns out that the NTIA is using the number of locations with speeds under 25/3 Mbps to allocate the $42.5 BEAD grant dollars between states. I happen to know that North Carolina has an aggressive mapping team that no doubt leaned on UScellular to fix its maps. But other states are probably not as aggressive and are losing a lot of BEAD grant dollars if they didn’t fix similar claims.

The problem is that, in many cases, the claimed speeds are not true. The technologies that I see claiming exactly 25/3 Mbps speeds include DSL, fixed cellular broadband, and fixed WISP broadband. It’s not hard to find examples where cellular companies and WISPs are claiming identical speeds across large geographic areas. That’s not how wireless technologies works. Speeds steadily decrease with the distance between a customer and a tower. Even if these ISPs are being truthful about the speeds that can be delivered close to towers, these speeds are exaggerated for customers farther from towers. It sounds like UScellular in North Carolina might have acknowledged the physics in lowering its claims.

It’s important to point out that any ISP claiming 25/3 Mbps is probably doing nothing wrong under the FCC mapping rules. The FCC allows ISP to report marketing speeds instead of trying to report accurate actual speeds. As long as an ISP is advertising speeds of ‘up to 25/3 Mbps’, it is not breaking the FCC mapping rules.

The title of this blog calls this a fixation on 25/3 Mbps. This particular problem was created when the politicians that wrote the BEAD grant (and other federal grants) rules tied eligibility for funding to the FCC maps and to specific speeds. This is an incredibly shortsighted idea because, by definition, the FCC’s allows ISPs to overstate speeds.

I have no doubt that tying grant eligibility to the FCC maps was done at the prompting of the biggest ISPs who undoubtedly helped to write the BEAD grant rules. The other speed used to define grant-eligible areas is 100/20  Mbps. That speed was chosen for the BEAD rules after a lot of political wrangling. The 100/20 Mbps speed was chosen to provide cover to cable companies to avoid the risk of being overbuilt by BEAD grants. Even if a cable company isn’t quite delivering an upload speeds of 20 Mbps, it can avoid entanglements with grants by declaring in the FCC maps it is meeting that speed.

It’s hard to understand the motivation of an ISP that claims 25/3 Mbps broadband when it knows it is delivering something slower. In this case, the overstatement of speed means that a state will get less BEAD grant funding – something the ISP might be interested in receiving. A claimed speed of 25/3 doesn’t make an area ineligible for grant funding – it just reduces the funding the state will receive. I suspect most ISPs claiming 25/3 Mbps are just declaring the marketing speed and don’t have any other hidden agenda.

While false claims of ISPs delivering 25/3 hurt the amount of BEAD funding coming to a state, the real problem comes from ISPs that falsely claim to be able to deliver 100/20 Mbps broadband. Under the BEAD and other federal and state grant rules, such areas are considered as served and grant money can’t be used to compete in these areas. All an ISP has to do to keep away grant competition is to claim a marketing speed of 100/20 Mbps. I think many of the ISPs making claiming 100/20 Mbps know exactly what they are doing – they are trying to fend off faster competition.

To me, this is akin to when CenturyLink and Frontier changed the claimed speeds in tens of thousands of Census blocks right before the FCC determined the areas eligible for RDOF. The FCC rejected those last-minute changes, which were blatant attempts to protect monopoly rural areas from the RDOF funding.

Just as an aside, the FCC supposedly was going to make RDOF available to all unserved locations in the country. Even with the above reporting issues that are badly suppressing the count of unserved locations, there are still 8.3 million unserved areas in the country, according to the latest FCC maps. The fact that RDOF missed 8.3 million eligible locations is probably the best example of why grants and subsidies should never be tied to the FCC broadband maps.

I have a simple fix for the current situation. If I was the NTIA, I would make one simple change before allocating BEAD dollars. I’d reduce every claimed speed of exactly 25 Mbps broadband to 24 Mbps. I’d then invite ISPs to prove they can meet the 25/3 speed.  I would do the same and reduce every claim of exactly 100 Mbps download to 99 Mbps when defining grant eligible areas. My guess is that this change would produce more accurate maps than the ones we have now. I’d give ISPs plenty of time to claim that the adjustment is wrong – which would be a better conversation than the flurry of map challenges we’re now seeing.

However, while this idea has appeal, it would be another rule based on our fixation of defining broadband with a specific speed number.

The Public Loves Fiber

The latest Customer Satisfaction Index is out from ACSI, which measures the public satisfaction of a wide range of U.S. industries and institutions. The survey this year continued to show that the public has a poor opinion of ISPs. As a group, ISPs had an average ACSI annual rating of 68. The only industry with a lower rating is gas stations at 65. Subscription TV had an average rating of 69, and the U.S. Post Office had a rating of 70.

But there is some interesting good news for some ISPs. Companies serving customers with fiber rated higher with the public than other ISPs, including cable companies using coaxial networks. Consider the following table that shows the 2023 ranking for fiber and non-fiber ISPs.

Fiber Non-Fiber
Altice 58
AT&T 80 72
Cable One 71
CenturyLink 78 62
Charter 64
Comcast 73 68
Cox 64
Frontier 74 61
Google Fiber 76
Mediacom 65
T-Mobile 73
Verizon 75
Windstream 70

For companies that offer both fiber and another technology, customers served by fiber liked an ISP more than non-fiber customers. CenturyLink has the biggest difference in satisfaction (78 for fiber and 62 for non-fiber). Frontier also has a dramatic difference (74 fiber and 61 non-fiber). The only cable company ranked for both technologies also has a sizeable difference, and Comcast has a ranking of 73 for its fiber network versus 68 for the coaxial network.

Customer satisfaction involves many other factors than just technology, but the differences for the companies that offer multiple technologies have to be mostly related to fiber. However, there are other factors in play. For example, it seems likely that CenturyLink and Frontier provide better customer service and faster repairs for fiber customers than for DSL customers.

Cable companies have to be noticing this giant difference as part of any consideration of how to upgrade their networks. The big cable companies are all at the beginning of the upgrades to improve upload speeds on coaxial networks, and they must be hoping that customers like them more after the upgrades. But there is a chance that the public has come to think of fiber as a superior technology and will not rank a coaxial system as highly even after speed increases. There is still a noticeable difference in latency and jitter between cable and fiber networks, and customers who see both in action believe fiber is better.

There is still a noticeable range of ISP rankings within each list. Non-fiber customers rate T-Mobile and AT&T the highest and rank Altice and Frontier DSL as the worst ISPs. It’s interesting to see Charter near the bottom of the rankings.

Fiber customers clearly rate AT&T as the best and Comcast Fiber as the lowest. Fiber technical performance should be consistent regardless of the ISP, so the difference in rankings between fiber providers has to be related to customer service and the other non-technical aspects of being an ISP.

Subsidizing Lost Cable TV Revenue

I’ve been tracking the number of claimed broadband customers at the largest telcos and cable companies for years. When I was looking at the statistics for the first quarter of 2023, it struck me that the biggest cable companies are now making up for the loss of cable TV customers by increasing broadband rates.

For the last decade, the big cable companies have been thriving financially through big annual increases in broadband customers. This came year after year as large numbers of customers bailed on DSL, along with organic population growth – the cable companies won most new broadband households.

Every traditional cable TV provider started to lose traditional cable TV customers starting around the end of 2018. Since that time, the loss of cable TV customers has accelerated. Small ISPs will tell you that they don’t make much money from cable TV, and many of them have abandoned the business line. But this is not the case for the biggest cable companies, which still have a decent monthly margin from cable TV.

But the margins on broadband are far higher than on cable TV, and as cable TV customers dropped, cable companies continued to add broadband customers, and profits continued to increase.

Profits were further bolstered by broadband rate increases. Comcast has been increasing broadband rates by about $3 per year, while Charter’s rate increases have been closer to $5. Charter broadband rates were significantly lower than Comcast, and they seem to be in the process of closing the gap.

Over the last year, the traditional formula of covering the losses of cable TV customers with the growth of broadband customers came to a screeching halt. Consider the following table that shows the annual change in broadband and cable TV customers for Comcast and Charter since the end of the first quarter of 2019. These numbers come from the Leichtman Research Group, which publishes customers every quarter for the biggest ISPs and cable providers.

Comcast Broadband Cable TV
YE 1Q 2020 1,509,000 (1,021,000)
YE 1Q 2021 1,928,000 (1,490,000)
YE 1Q 2022 1,129,000 (1,691,000)
YE 1Q 2023 161,000 (2,136,000)
Charter Broadband Cable TV
YE 1Q 2020 1,559,000 (357,000)
YE 1Q 2021 1,988,000 (12,000)
YE 1Q 2022 1,040,000 (341,000)
YE 1Q 2023 235,000 (815,000)

In the years ended in the first quarter of 2020 and 2021, Comcast added more broadband customers than the losses of cable TV customers. That flipped in the year ended 1Q 2022 as Comcast lost 562,000 more cable TV customers than it gained broadband customers. In the year just ended 1Q 2023, the wheels have totally come off for Comcast. The company lost over 2.1 million cable TV customers while gaining only 161,000 broadband customers.

Charter’s numbers are not quite as dramatic since the company has been able to hang onto traditional cable TV customers better than the rest of the industry. But Charter’s net customer gains for broadband have slowed to only 235,000 for the year ended 1Q 2023 while traditional cable TV losses are accelerating.

It now seems like both big cable companies must make up for losses of cable TV customers through broadband rate increases. The day of broadband growth subsidizing cable TV losses is over. I wonder how folks who cut the cord from these two companies feel about having the companies make up for losing them as cable TV customers by raising broadband rates?