Blog

Pinocchio

Unlimited Plans At 2G Speeds Are Bogus

Christian 12 Comments

It’s becoming more common for carriers to offer additional data at 2G speeds after subscribers use up all of the regular-speed data that they’ve been allotted. In most cases, this means subscribers who’ve run out of regular data are throttled to a maximum speed of 128Kbps. It’s a great perk. Imagine you’ve run out of regular data, but really need to use the internet for a moment to pull up a boarding pass, look up directions, or view an email. At 2G speeds, it will probably be frustratingly slow to do any of those things, but that’s a much better scenario than being unable to use data at all.

Most consumers have little clue what 2G speeds amount to in practice. Let me be clear: 2G speeds are really slow for most things people want to do. Music streaming probably won’t work well. Video streaming at low, 240p resolution won’t be possible. Most websites will take a long time to load.

Carriers vary in how they present the perk of extra data at 2G speeds. In my opinion, Mint Mobile and Verizon handle the perk in a commendable way. Both carriers generally describe their plans and data allotments based on the amount of regular data allotted. In contrast, Total Wireless and Tello offer “unlimited” plans. These plans have caps on regular data use. After the cap is reached, subscribers continue to have data at 2G speeds. I think it’s misleading, bordering on outright lying, to call these unlimited plans. It’s just not possible to use data in a normal manner once speeds are throttled to 128Kbps.

In fact, imposing a throttle creates a limit on how much data can be used in a month. If a subscriber manages to transmit 128 kilobits of data every second for an entire month, they’ll use about 40GB of data.1 While almost no subscribers will come close to reaching it, there is a theoretical limit on these supposedly unlimited plans. It’s roughly: amount of regular data + 40GB.

Disclosure: I have financial relationships with Verizon, Mint Mobile, Tello, and Total Wireless (more details).

T-Mobile Launches New Test Drive Program

Christian

In the last couple of days, T-Mobile has begun a new trial program. Non-customers can get a mobile hotspot along with 30 days of service and 30GB of data at no cost.

The program provides an easy option for people to test how well T-Mobile’s network could work for them. Once a hotspot is turned on, users can connect their existing phones and devices via Wi-Fi. The hotspots are compatible with LTE band 71, so hotspot users can experience the benefits of T-Mobile’s new Extended Range LTE signal.1

What’s the catch? As far as I can tell, there isn’t one. Signing up is easy, no credit card is required, and it’s not even necessary to return the hotspot at the end of the trial. I’ve joined the program, and I’ll probably post again in about a month with an update on my experience.

You can sign up for the program here. T-Mobile’s CEO, John Legere, discusses the program in more detail in the video below.

Unihertz Atom Review – A Postmortem

Christian

If you happen to stumble upon a tiny phone on the bottom of the San Marcos river, please let me know. It’s mine.

Towards the end of last year, I supported the Indiegogo campaign for the Unihertz Atom. The Atom is a tiny, ultra-durable phone. I have a habit of putting my personal phones through hell, and I hate how large most of today’s mainstream phones are. I thought the Atom might be just what I needed.

I used the Atom for something like nine months before losing it this weekend. The form factor of the phone is probably the most interesting thing about it:1

The phone is thick but otherwise tiny. Before buying the phone, I didn’t realize how much of a conversation piece it would become. If you buy one, prepare yourself for endless questions along the lines of “Is that a [pause] phone?” and “How do you text on that?”

Performance

I’m not going to dive into details about the phone’s hardware specs. Plenty of other reviewers have already done that. At a high-level, the phone has decent hardware given its small size and low price (about $250). I never had any trouble with sluggish performance. After all, given the phone’s small size, it’s not too tempting to multitask aggressively or use intensely demanding apps in the same way you might on a conventional phone.

Size constraints

Most apps were surprisingly good at accommodating the Atom’s small screen. Texting and emailing weren’t as pleasant as they would be on a larger device, but neither activity involved a lot of struggling. Android’s auto-correct features were pretty useful for keeping the typing experience fluid despite occasionally hitting the wrong letters on the keyboard. That said, I have small fingers and good eyes. Other people might find typing more painful than I did.

Despite the phone’s size, it still has basically everything I expect a normal phone to have. The Atom has a flashlight. There’s a rear camera and a front-facing camera. Both cameras are lousy, but they work.

Overall, I think most aspects of the phone can be described as: passable but not great. That’s sort of the point of the Atom. You can use the Atom to waste time on social media, but it’s not as pleasant as the same time wasting would be on another phone.

While using the Atom, I found myself being less responsive to messages than I usually am. Reduced responsiveness could be a good thing if you’re frustrated with how attached you are to your phone. In my case, I think reduced responsiveness was a bad thing.

Durability

Unihertz claims the Atom is IP68 rated, meaning that it is both dust-proof and waterproof. Neither the headphone jack or the charging port are sealed, so I was initially afraid to put the phone to the test. Once I got over the fear, my Atom survived plenty of time underwater.

I’m not sure how it happened, but I did manage to put a decent scratch into the Atom’s screen at some point. I don’t think the screen is especially fragile, but it’s less scratch-resistant than sapphire screens seen on some rugged phones.

Other issues

  • Call quality often seemed bad. I don’t think this was so much about the phone being unable to offer good call quality as it was about me struggling to position the phone so that my ear was by the speaker while the microphone was well-placed.
  • The Atom’s vibrate function is weak. I wouldn’t always feel it in my pocket. I see this as a substantial negative. I don’t like putting my phone on ring.
  • Unihertz’s communication between the time when I supported the Indiegogo campaign and the time when I received the device was not very good. The phone also shipped slowly. I would have got my device sooner if I had just purchased an Atom via Amazon after the release date.

Other positives

  • Unihertz has some accessories (an armband, a bike mount, a clip mount) built specifically for the Atom. Rather than charging a premium for these accessories since normal accessories won’t fit the Atom, Unihertz sells them for entirely reasonable prices in the $10-$20 range.
  • The Atom works with AT&T, Verizon, and T-Mobile. However, it doesn’t qualify as what I call a nearly universal unlocked phone since it lacks whitelisting from Sprint and does not have complete support for all the important LTE bands used in the U.S.
  • The Atom has a dual SIM tray. You can attach multiple lines of service to a single device.
  • The Atom’s battery life is excellent. This may be more about the Atom being unappealing for frequent use than the battery capacity being especially good.

Closing thoughts

Would I recommend the Atom? I’m not sure. It was a fun experiment using it for most of a year. If you’re weird in the same sorts of ways I am, you’ll enjoy a lot of things about the Atom. It’s certainly durable, it doesn’t take up a lot of room in pockets, and it’s great for bringing along on bike rides. That said, the Atom was sometimes a pain to use as a primary phone. When I lost the phone while tubing down a river last week, I may have been more relieved than upset.

Avoid Mismatched Phone Plans

Christian

There are probably millions of people in the U.S. that could save a lot of money by switching to a different plan offered by their existing cell phone carrier. For example, plenty of people pay for expensive plans with unlimited data, even though they only use a few gigabytes of data each month.

Recently, I angered a lot of people when I said Google Fi is generally too expensive for me to recommend the service. Several commenters argued I was wrong. Some of the commenters were polite. Others called me an idiot. Commenters often mentioned how much they used to pay for service from a major carrier and how much they saved by switching to Google Fi.

In many cases, commenters appeared to have purchased the wrong plans when they used major carriers. They were paying for data they didn’t need. Since Google Fi has a pay-for-what-you-use pricing structure, Fi subscribers basically cannot be on a plan that is mismatched with their data use.

Unsurprisingly, a person who barely uses data can probably get a better rate on a Google Fi plan than she can get on a high-data plan from Verizon. But Verizon also offers plans with small data allotments. We should make apples-to-apples comparisons when we can.

Examples

Below, I share excerpts from previous comments and my follow-up thoughts.

My wireless bill with Fi is $100 cheaper than it was with two phones on Verizon’s cheapest plan…a plan that includes more than 1gb per a phone is costly and unnecessary.
$100 cheaper!? I don’t think this commenter could have been on Verizon’s cheapest plan. Today, two lines of Verizon prepaid with 6 GB of data on each line (way more than the commenter desires) would cost only a bit more than $60 per month.

Google Fi unlimited calls and texts only costs $20 a month and when you add that to their pay-for-what-you-use data your monthly cost could be around mine at roughly $28/month, as I barely use any data…Now compare that with Verizon’s bare minimum unlimited plan starting at $70 before taxes and fees…Fi allows us to escape the tyranny of major cellular corporations and their overpriced plan structures.1
No! It’s inappropriate to compare the cost of service with barely any data use to the cost of an unlimited data plan.

Our monthly bill for all 3 lines with Verizon was around $180. It was reduced to less than $70 after switching to FI for the last 5 months.
Under $70 for three lines is a pretty good deal! No need to switch away from Fi, but let’s consider what comparable service would cost today with Verizon. With three Fi lines and a total cost under $70 per month, total data use is probably under 2GB per month.2 A postpaid, Verizon plan with 3 lines and 2GB of shared data is about $100 per month right now. Prepaid options could come out under $100 per month.

My bill with Verizon was always $105 a month for two gigs of Internet.
One of Verizon’s prepaid options right now offers three times that amount of data for about a third of the price!

Carriers create confusion

People who are on mismatched plans aren’t idiots. Many carriers like it went customers pay extra money for unnecessary amounts of data. Instead of alerting subscribers who are paying for too much data, carriers often take steps to encourage customers to over-purchase data. I call the cell phone industry a confusopoly for a reason.

Finding plans that fit

As mentioned earlier, one way to ensure that you’re not paying for data you don’t need is to choose a carrier with a pay-for-what-you-use model (e.g., Ting or Google Fi). That said, I think most people can find better prices with carriers that use conventional pricing structures.

If you know how much data you typically use (or have records of data use you can look back on), you can probably figure out how much data you’d like your cell phone plan to offer. If you’re unsure about your data use, I suggest starting small. Choose a plan with the smallest amount of data that you think might be adequate. Experiment with that plan for a few months. Add more data if the initial data allotment you started with turns out to be insufficient.

Did Google Fi Shoot Itself In The Foot?

Christian 39 Comments

Google Fi has a lot going for it: amazing international roaming options, fancy network-switching technology, and a simple pricing structure. Despite all Fi’s great aspects, I don’t usually recommend it. For most users, it’s just too expensive. Google Fi typically charges $10 per gigabyte of data. A lot of other carriers offer plans with far lower rates for data.

All Fi subscribers have roughly the same plan with the same pricing structure.1 There aren’t ten different plans with different names and policies. This is in sharp contrast with Verizon. Looking at just unlimited plans, Verizon has several options:

  1. Start Unlimited
  2. Play More Unlimited
  3. Do More Unlimited
  4. Get More Unlimited

In fact, Verizon actually has a fifth unlimited plan it offers as a prepaid option. Each unlimited plan is a bit different. Some of the plans have more limits than others—inviting critics to joke about how Verizon doesn’t understand the meaning of the word “unlimited.”

While it feels silly, there are a handful of reasons why it makes business sense for Verizon to have several unlimited plans. Today, I’ll only touch on one of those reasons: when a carrier has multiple plans, it’s easier to introduce new prices and policies without immediately affecting existing customers. We just saw Verizon do this. A month ago, Verizon was offering three postpaid, unlimited plans. They were different from today’s plans:

  • GoUnlimited
  • BeyondUnlimited
  • AboveUnlimited

When Verizon introduces new plans, it can cease offering old plans to new customers while offering existing customers the same service on legacy plans. Since there are several plans that all have different policies, it’s difficult for people to make simple, apples-to-apples comparisons between legacy plans and plans available to new customers.

Back to Fi. Google Fi has been charging almost everyone $10 per gigabyte for a long time.2 Years ago, that was a decent price for data. Today it’s not. Data costs have gone down in most of the industry.

I don’t have any inside knowledge about Fi, but I’m suspicious Fi’s simple pricing structure makes it hard for the company to change its prices. If Fi wanted to offer new customers data for $5 per gigabyte, existing Google Fi subscribers would want that deal too. If existing subscribers had to continue paying $10 per gigabyte, they’d get angry. If Fi reduced prices for existing subscribers, Fi’s revenue would plummet.

Added after publication: The idea I share in this post probably doesn’t explain why Fi charges so much for data (or at least, it is probably an incomplete explanation). There are a lot of other plausible explanations. E.g., Fi’s agreements with network operators may not lead to Fi getting good rates on data.

Added even later: When I said I don’t usually recommend Google Fi, I didn’t mean to imply that Fi’s prices are uniquely awful or that no one should use Fi. Rather, I don’t typically recommend Google Fi since most consumers can find comparable service at a lower price (see carriers I recommend).

Start Small

Christian

With recurring expenses, I often advise people to start small and upgrade later if necessary. The rationale behind the advice is easy to illustrate with internet service.

Home internet is often priced based on the max speeds the service provider will deliver. Usually, max speeds are described as some number of megabits per second (Mbps). A cable company’s price structure might look something like this:

  • 40 Mbps – $30 per month
  • 150 Mbps – $45 per month
  • 300 Mbps – $60 per month
  • 600 Mbps – $75 per month
  • 1,000 Mbps – $100 per month

The problem

Most people want internet that feels fast, but not many people have a clear sense of how many Mbps it takes for a connection to feel fast. People who are uncertain often end up choosing a speed that falls in the middle of the options available. If that speed is sufficient, they generally stick with it. If the speed turns out to be too slow, they upgrade. With this approach, people won’t get clear feedback if they purchase faster speeds than necessary. People often spend years paying extra for high speeds they don’t benefit from.

In my opinion, over-purchasing happens more often than necessary because service providers encourage it. For example, one of the Xfinity internet options available to me right now involves a max speed of 60 Mbps. Xfinity explains that the speed is “good for up to 5 devices at the same time.” This is silly. 60Mbps might only support 5 devices if they’re all streaming ultra-HD video at the same time. For more realistic situations, a 60 Mbps connection can support far more than 5 devices. Xfinity knows this, but Xfinity has an incentive to encourage customers to purchase more expensive service than necessary.

I expect Xfinity’s cost structure is effective for the company since it allows them to engage in a weak form of price discrimination. In most markets, different people are willing to pay different amounts for the same product or service. If business owners can find a way to charge more money to the people who are willing to pay more, their businesses will be more profitable. Something like this occurs with internet service. People with tight budgets tend to start with cheaper, slower options. People with more money tend to purchase higher speeds than they need and often overpay for service without recognizing it.

The solution

Overspending is often easy to avoid. When services allow easy upgrading without any extra fees (as is often the case with internet), I advise people to start with the cheapest option that they think might be adequate. If you start with a low-speed tier for internet services, there’s a good chance you’ll find it satisfactory. If not, you can upgrade to the next speed tier.

The same solution works in other industries. Not sure how many gigabytes of data you need on your cell phone plan? Start with a small amount. If you hit your monthly allotment, add more data.

Not sure whether cheap, prepaid service will perform as well as postpaid service? Try prepaid service for a month. If you like it, stick with it. If not, switch over to postpaid service.

Starting small is a good idea in situations where services with recurring bills are easy to upgrade. While companies often penalize people who downgrade services, upgrading services is often easier. Companies are generally happy to have their customers pay more each month. That said, starting small isn’t always a good idea. For example, an internet service provider may offer discounted, introductory rates that customers are ineligible for when upgrading.

Buy Your Own Router And Modem

Christian

Earlier this week, I signed up for Comcast’s Xfinity internet at a new apartment. If a subscriber doesn’t bring his or her own modem and router, Xfinity will rent a device for $13 per month. As far as I can tell, the majority of subscribers opt to rent. Renting is a terrible deal.

I purchased a modem and a router in 2017 for a total of about $70. While I don’t mind having a separate modem and router, there are plenty of options for simple devices that combine a router and a modem into a single unit. Amazon’s best-selling combination devices can be seen here. At the time of writing, several high-quality devices are available for between $80 and $130.

A $13 per month rental fee works out to $156 per year. A router can easily be used for 3 years, possibly much longer.

The table below shows the overall cost of renting a $13 per month router/modem for different lengths of time. I assume that rental fees will not increase over time. The assumption is generous. In the last decade, Comcast’s rental costs have risen by about 4X from $3 per month to $13 per month. The table also shows how much a subscriber could save by buying a high-end, $150 router/modem instead of renting.

Years usedRental costSavings by purchasingRental cost vs. purchase cost
1$156$64%
2$312$162108%
3$468$318212%
4$624$474316%
5$780$630420%
6$936$786524%
7$1,092$942628%

If you only use a $150 device for a single year, renting and buying are about equally cost-effective. In the off chance you use your service for less than a year, renting may make financial sense. That said, most Xfinity plans come with a one-year commitment. I doubt many customers subscribe for short enough periods to justify renting. If you keep your service for a few years or more, you can save a ton of money by buying your own modem and router.

Renting a modem/router may appeal to people since it’s simple. If you rent a device from Xfinity, you may not worry about compatibility and performance as much as you would when buying your own modem/router.

If the technical aspects of buying a modem or router concern you, here’s my advice for keeping things simple:

  • Visit Amazon’s page listing its best-selling modem/router combination devices.
  • Limit yourself to devices that are well-reviewed.
  • Find a listing explicitly states that the device is (a) compatible with the service you will use (e.g., Comcast Xfinity) and (b) able to support service at the speed you want.1

Some people will struggle when deciding which device to purchase. A lot of consumers aren’t sure about the speeds they’ll want. They might find a cheap device that is probably good enough but wonder whether they should spend some extra cash to get a device they’re confident is good enough. Remember that even if you spend more than you need to on a router/modem, you’re probably still getting a way better deal than you would with a rented device.

2,000,000 AT&T Phones Were Unlocked Illegally

Christian

It recently came out that around 2,000,000 AT&T phones were unlocked by hackers that bribed AT&T employees. Muhammad Fahd and co-conspirators allegedly bribed a handful of AT&T employees to make the unlocks possible.

As I understand it, around 2012 lists of IMEI numbers were provided to bribed employees so that devices could be fraudulently unlocked. Eventually, the crimes became more involved. Bribed employees installed malware on AT&T systems and fraudulent wireless access points in AT&T facilities.

It’s a crazy story. Several years ago, I wondered how so many third-party services managed to offer device unlocking. I suppose this story is part of the explanation.

For more details, check out Ars Technica’s article.

Photo of a frustrated person with a broken phone

Consumer Reports’ Broken Cell Service Rankings

Christian

Several months ago, I published a blog post arguing that Consumer Reports’ cell phone rankings were broken. This month, Consumer Reports updated those rankings with data from another round of surveying its subscribers. The rankings are still broken.

Consumer Reports slightly changed its approach this round. While Consumer Reports used to share results on 7 metrics, it now uses 5 metrics:

  1. Value
  2. Customer support
  3. Data
  4. Reception
  5. Telemarketing call frequency

Of the 19 carriers Consumer Reports’ assesses, only 5 operate their own network hardware.1 The other 14 carriers resell access to other companies’ networks while maintaining their own customer support teams and retail presences.2

Several of the carriers that don’t run their own network offer service over only one host network:

  • Cricket Wireless – AT&T’s network
  • Page Plus Cellular – Verizon’s network
  • MetroPCS – T-Mobile’s network
  • CREDO Mobile – Verizon’s network
  • Boost Mobile – Sprint’s network
  • GreatCall – Verizon’s network
  • Virgin Mobile – Sprint’s network

To test the validity of Consumer Reports’ methodology, we can compare scores on metrics assessing network quality between each of these carriers and their host network. At first glance, it looks like the reception and data metrics should both be exclusively about network quality. However, the scores for data account for value as well as quality:3

Data service indicates overall experience (e.g., cost, speed, reliability) with the data service.
I think it was a methodological mistake to account for value within the data metric then account for value again in the value metric. That leaves us with only the reception scores.4 Here are the scores the four host operators get for reception:

  • Verizon – Good
  • T-Mobile – Fair
  • AT&T – Poor
  • Sprint – Poor

How do those companies’ scores compare to scores earned by carriers that piggyback on their networks?

  • Cricket Wireless has good reception while AT&T has poor reception.
  • Page Plus and Verizon both have good reception.
  • MetroPCS has good reception while T-Mobile has fair reception.
  • CREDO and Verizon both have good reception.
  • Boost has very good reception while Sprint has poor reception.
  • GreatCall and Verizon both have good reception.
  • Virgin has good reception while Sprint has poor reception.

In the majority of cases, carriers beat their host networks. The massive differences between Cricket/AT&T and Boost/Sprint are especially concerning. In no cases do host operators beat the carriers that piggyback on their networks. I would have expected the opposite outcome. Host networks generally give higher priority to their direct subscribers when networks are busy.

The rankings are broken.

What’s the problem?

I see two especially plausible explanations for why the survey results aren’t valid for comparison purposes:

  • Non-independent scoring – Respondents may take prices into account when assessing metrics other than value. If that happens, scores won’t be valid for comparisons across carriers.
  • Selection bias – Respondents were not randomly selected to try certain carriers. Accordingly, respondents who use a given carrier probably differ systematically from respondents that use another carrier. Differences in scores between two carriers could reflect either (a) genuine differences in service quality or (b) differences in the type of people who use each service.

Consumer Reports, please do better!

My earlier blog post about Consumer Reports’ methodology is one of the most popular articles I’ve written. I’m nearly certain staff at Consumer Reports have read it. I’ve tried to reach out to Consumer Reports through two different channels. First, I was ignored. Later, I got a response indicating that an editor might reach out to me. So far, that hasn’t happened.

I see three reasonable ways for Consumer Reports’ to respond to the issues I’ve raised:

  • Adjust the survey methodology.
  • Cease ranking cell phone carriers.
  • Continue with the existing methodology, but mention its serious problems prominently when discussing results.

Continuing to publishing rankings based on a broken methodology without disclosing problems is irresponsible.

Markets Are Honest

Christian

I’ve been reading a ton of articles with commentators’ takes on whether a merger between Sprint and T-Mobile will be good or bad for consumers. Almost everything I’ve read has taken a strong position one way or the other. I don’t think I’ve seen a single article that expressed substantial uncertainty about whether a merger would be good or bad.

It could be that everyone is hugely biased on both sides of the argument. Or maybe the deal is so bad that only incredibly biased people would consider making an argument that the merger will be good for consumers. I’m not sure.

I like to look at how markets handle situations I’m uncertain about. In the last few years, I’ve regularly seen liberal politicians and liberal news agencies arguing that we’re about to see the end of Trump’s presidency because of some supposedly impeachable action that just came to light. I’m not Trump’s biggest fan, but I’ve found a lot of arguments about how he’s about to be impeached too far-fetched. I have a habit of going to the political betting market PredictIt when I see new arguments of this sort. PredictIt has markets on lots of topics, including whether or not Trump will be impeached.

Politicians and newspapers have an incentive to say things that will generate attention. A lot of the time, doing what gets attention is at odds with saying what’s true. People putting money in markets have incentives that are better aligned with truth.

Most of the time I’ve seen articles about Trump’s impending impeachment, political betting markets haven’t moved much. In rare occasions where markets moved significantly, I’ve had a good indication that something major actually happened.

Wall Street investors have a strong incentive to understand how the merger will actually affect network operators’ success. Unsurprisingly, T-Mobile’s stock increased substantially when key information indicating likely approval of a merger came out. Sprint’s stock also increased in value.

What’s much weirder is that neither Verizon’s stock nor AT&T’s stock seemed to take a negative hit on the days when important information about the merger’s likelihood came out. In fact, it actually looks like the stocks may have increased slightly in value.1

You could tell complicated stories to explain why a merger could be good for competing companies’ stock prices and also good for consumers. I think the simpler story is much more plausible: Wall Street is betting the merger will be bad for consumers.

Maybe none of this should be surprising. There were other honest signals earlier on in the approval process. As far as I can tell, neither Verizon nor AT&T seriously resisted the merger:2

Disclosure: At the time of writing, I have financial relationships with a bunch of telecommunications companies, including all of the major U.S. network operators except T-Mobile.