With apologies to any Marines, the question around this post is whether Project Fi is ready for commitment. For those unaware, Semper Fi stands for Semper Fidelis, latin for “always faithful.” You may recognize it as the Marine Corp motto. For those even more unaware, Project Fi is Google’s attempt at creating a way to access mobile voice and data outside of the tradition carriers such as AT&T and Verizon by creating its own “carrier,” traditionally described as an MVNO (Mobile Virtual Network Operator).
Unlike MVNOs of the past that only supported one carrier or other carriers such as T-Mobile trying to blend WiFi and cellular, Google is trying to blend service from Sprint, T-Mobile, and WiFi into a seamless experience. They are rolling it out very slowly, but I was lucky to get an early invite. I’ve been using it in parallel with my AT&T phone over the last few weeks to measure performance and reliability across Nashville’s tricky-for-wireless-signals topography.
To begin, let us look at what they sent me. Google, like many other companies, is striving to meet the challenge created by Apple’s excellent packaging. In addition to a small package containing the SIM care, they also sent a large box that involved suspense. At launch, Project Fi only supports the Nexus 6. Since this was only a test for me, I chose to buy a used one instead of purchasing a new one from Google. So why did they ship me this big box? And what could be in it? In one word: swag.
But good swag! A decent set of earbuds, a 6000mAh battery, and a nice case. All of it is completely irrelevant to using Project Fi, but the box is a nice quality touch that endears the user immediately. The package with the SIM card even came with a paper clip for easy card installation.
According to internet chatter, the intro boxes are not always shipping on time and the rumor is that these will stop at some point. This is a shame, because what followed next needed some serious goodwill.
I want to avoid a long and dull description of what it took to get the phone, SIM, and Project Fi running, so here is the summary: painful. If it was a problem for me, someone who is technically competent and relatively patient with glitches, I can’t imagine what it would be like for other, less tech-savvy customers. Some of it may have been my fault for not reading instructions, and some of it may have been the used phone, but if it takes two system updates and two restarts as a normal procedure, Project Fi is dead. However, since they are dealing with new firmware, software, and SIM technology, it will likely take a while before the on-boarding procedure is as easy as dropping in a new SIM.
The next item to discuss is the impact Project Fi will have on your Google Voice number. I’ve had a Google Voice number since the first month they were offered. It is essentially a VOIP number that can be used in Hangouts or forwarded to another number. I use it as a work number that will still ring my personal phone, giving me a pseudo dual-SIM setup on a regular phone. It has been indispensable when dealing with people who I do not want to have my personal number. The only downside is the confusion when I respond to them, inadvertently, with my personal number when calling or texting. Texting with my Google Voice number requires me to use the Google Voice app, so it is not part of my normal workflow.
Project Fi uses much of the Google Voice infrastructure for messaging, voice mail, and hangouts, which is good but also forces a choice between three options when you adopt Project Fi.
Option 1: Port your existing carrier phone number over and it will replace your current Google Voice number, returning that GV number to the Google Voice pool for other users to claim. This is not a good option for me, as I have used my Google Voice number as a business number for over five years. Losing this number that is on business cards and in email signatures would not only be disruptive, it would force me to use my personal cell number for all my business dealings.
Option 2: Keep your current Google Voice number as the Project Fi number. End your current carrier contract and lose that number. This is even more repugnant to me, as I am unwilling to give up my personal number that I have had for sixteen years.
Option 3: Create another Google account for your Project Fi invite. That way you can keep your existing Google Voice number and use your new Google account GV number for Project Fi, with the added complexity of managing yet another Google account. Unfortunately, I found out about this option too late and did not want to start all over again at the end of the Project Fi invite queue.
The remainder of this post will detail some of my actual experience in using Project Fi.
My Project Fi phone was first activated at home. My neighborhood is notorious for bad cell reception, and I rely on a AT&T microcell to have reliable AT&T service in my house. In fact, that is one of the reasons I was interested in Project Fi. I would love to rid myself of the microcell and not have to worry about handoffs when arriving or leaving my house.
As mentioned previously, Project Fi uses a combination of Wifi, T-Mobile cellular, and Sprint cellular signals to provide voice and data services. The phone worked great on my Wifi, and placing calls was easy, reliable, and consistent. I turned off the Wifi in anxious anticipation of what cellular signal strength I would get from Fi. After a short 5-10 second delay, up popped a T-Mobile LTE signal that, while not 5 bar strength, gave me great coverage in all parts of my house. Wonderful!
Next, I took our little Chihuahua for our normal walk around the neighborhood. Uh oh. Two doors down the T-Mobile signal disappeared and a Sprint signal showed up, but with a graphic I would soon come to hate, the dreaded exclamation mark on the signal strength icon:
So, there I was with no data signal through most of my walk. Subsequent journeys to other parts of Nashville showed the same tendency – carrier signal, but no data connectivity.
I resisted my normal, geeky tendency to search on forums and instead tried out the Project Fi help line. I was pleasantly surprised by a polite, understanding, and knowledgeable person on the other end of the phone. He led me through a few basic questions before having me key in a code on my dialer, *#*#346398#*#*. To those of us used to hacking phones, I recognized this type of code as a way to communicate with the firmware of the phone, accessing settings that are not available through the normal user interface. This code did work, usually taking me from a Sprint signal to a T-Mobile signal. I later did some research and found that this was a code to force the phone to jump to the next carrier. A quick search of the web found other codes that you can use to troubleshoot:
*#*#34777#*#* Force jump to Sprint
*#*#34866#*#* Force jump to T-Mobile
*#*#346398#*#* Force jump to the next carrier
*#*#342886#*#* Re-enable automatic carrier switching
(For those familiar with phone keypads, the number codes make even more sense if you see them in text: FISPR, FITMO, FINEXT, FIAUTO)
I headed to work on Monday to more in-depth testing. The ATT network is somewhat flakey at my work location as well, though since we’ve moved to the front of the building I haven’t had any major connection issues. Comparing my ATT phone and the two Project Fi carriers, I found good connections with all three. Speed was another issue. Of the three carriers, T-Mobile had the fastest speeds, something that I have heard from several people. Sprint had the worst, though I’ll talk about that later. ATT was still serviceable. Here are the Mbps speed numbers for download/upload: 7.23/.33 ATT LTE, 1.65/.78 PFi (Sprint EVDO), 10.33/9.82 PFi (T-Mobile LTE).
Throughout the next week around Nashville, this trend continued, with ATT and T-Mobile having similar coverage with Sprint having smaller coverage while seemingly limited to EVDO, not their LTE network. So why no LTE? No one seems to know, including the aforementioned great Project Fi tech support. This underscores the issue with Project Fi – it is truly a beta product, with transferring calls and signals between cell towers of different systems still being an unexplored territory.
The other large problem, unfortunately without hard data to support, is that both Sprint and T-Mobile are not giving Google full access to their network. While it makes sense that carriers would not want to penalize their own customers, the actual logic of throttling or limiting would be something outside of Google’s understanding or control. Many users have posted speed tests of Project Fi and T-Mobile phones side by side, connected to the same tower but with the T-Mobile phone having twice the speed. Others, myself included, have seen a disturbing lack of Sprint LTE connections where there should be plenty.
At the end of the first week, my family and I took a quick trip to Paducah, Kentucky. This gave me two more data points to consider. First, on the way up I-24 I had a very important call dropped. This is an example of not being able to trust the network and also not being willing to test it out. It could be that the call would have dropped even if I was using my AT&T phone. But since I can not test concurrently, I attribute the failure to Project Fi. Additionally, I also discovered that while on the call, I could not use data. This problem is well documented for Sprint LTE, but I seemed to be on T-Mobile service. You can see from the screenshot that, unfortunately, the signal had dropped down to GSM, eliminating the ability to use voice and data concurrently.
After using AT&T for so long with concurrent data, this would be a very hard thing to accept. Second, upon arriving in our Paducah Airbnb in Lowertown, I quickly discovered that I had zero coverage on Project Fi. No T-Mobile and no Sprint. This was the final, perhaps defeating blow. If I’m in a relatively large town, I expect coverage. The coverage maps for both Sprint and T-Mobile show LTE coverage, and the T-Mobile map even shows Customer Verified on the exact spot I was testing.
As you may have figured out by now, the answer to my opening question is no. Project Fi is not reliable enough right now to deserve my faithfulness. I know enough both technically and experientially to realize that what they are trying to do is very hard. Switching between three sources of signals (WiFi, Sprint, T-Mobile) compounded by multiple types of signal (GSM, EDGE, EVDO, LTE, etc.) per source is incredibly complex. However, in talks with their tech support I also know that Google is best suited to figure it out. Why? Because so much of this problem is signal data collection against a geographical data background. Google’s experience with Maps and Streetview gives them an advantage over other parties that might try to do the same thing. Over time, they will be able to optimize the hand-offs based on location, direction, speed, equipment, time of day, season, and weather. The big question is if that optimized hand-off network will ever be as reliable as other, single-sourced networks. Only time and experience will tell.
I want to believe in Project Fi. Just as Apple and the iPhone paved the way for carriers to become dumb pipes, I see Google as paving the way for wireless data to be a commodity, much like electricity and water. The Project Fi pricing is already putting pressure on the major players, joining T-Mobile and Sprint to help erode the contracts and subsidies that have ruled the U.S. carrier ecosystem for so long. Verizon is the latest carrier to announce a move away from contracts and to more simplified pricing. In the long term, this will be good for all of us.
Oh, and what Google Voice option did I choose? Actually a fourth option. I have kept my old Google Voice number as my Project Fi number and have kept my personal number with AT&T. I have decided to keep both, paying $20/month to Project Fi for carrier and connection diversity. Swapping SIMs back and forth is easy and gives me the option of moving over to Project Fi permanently if they fix the reliability issues.