T-Mobile Finally Sells Sprint Fiber

An Outside Perspective Overcomes Long-Held Cognitive Biases

According to Fierce Wireless, T-Mobile has agreed to sell the wireline business the company gained as part of its acquisition of Sprint to Cogent Communications for $1. In 2004, when I was director of strategic planning for the division of Sprint that managed the wireline business, we first seriously looked at selling this asset. Over the next 10 years in that role and as vice president of corporate strategy, we considered the possibility of selling Wireline several more times, even going so far as holding exploratory conversations with potential buyers.

The Sprint Fiber Optic Network

Each time we backed away, ostensibly because the business and the underlying assets were too essential to Sprint’s overall business. However, I think we simply suffered from some of the most common cognitive biases that lead many businesses to make bad decisions. Over the past 18 years several things have changed that enabled T-Mobile to finally make the right decision.

For starters, the wireline industry has continued to decline, and Sprint/T-Mobile’s wireline business with it. So, the wireline business, which once had been the largest and most profitable part of Sprint, had become a tiny and unprofitable part of T-Mobile. Second, the powerful individuals who strongly identified with the wireline business and challenged any attempts to divest it have moved on. Finally, and probably most importantly, I’m guessing T-Mobile’s leadership brought an outsider’s perspective that was able to cut through cognitive biases and make the right decision.

A Little History

By the 1970s, United Telecom was the 3rd largest telephone company in the U.S., providing telephone service to homes and businesses in about 5% of the country. Like virtually all other local telephone companies, the company relied on AT&T Long Lines to deliver long distance telephone calls. That began to change with the government’s breakup of AT&T in the early 1980s. United Telecom started to get into long distance services and, in 1984, announced plans to build a nationwide fiber optic network. The long distance business and fiber network were called US Sprint. By 1990 this had clearly become the core of the company. United Telecom changed its name to Sprint Corporation and changed its stock ticker to FON (for Fiber Optic Network). (You can read a more detailed history here.)

In the mid 1990s, Sprint saw another emerging opportunity — wireless. The company spent about $4 billion for nationwide spectrum licenses and another $10 billion to build out a nationwide network. By the time I arrived at Sprint in 2003, wireless and long distance wireline were about the same size, but wireless was on the rise and wireline was in decline. In 2004 we faced a crisis. The long distance wireline business had been generating about $1 billion each year in cash, which helped fund continued growth in wireless, but because of the overall decline in the long distance industry, we projected that cash flow would go negative within about 3 years. We needed to do something.

In broad general terms, we explored four options:

1. Do Nothing: This is always an option, but when facing crisis, usually not a good one.
2. Double Down: Buy either MCI or AT&T.
3. Exit: Sell the wireline business to Level 3, Qwest, or Global Crossing.
4. Prune: Exit portions of the business in order to enable other portions to thrive.

As we evaluated these options, we knew that doing nothing (option 1) wasn’t a viable option and doubling down (option 2) wouldn’t address the fundamental trends causing our problems. Which left exiting (option 3) and pruning (option 4). (More details here.)

(As a side note, the original United Telecom local telephone business was spun off as a separate entity called Embarq when Sprint and Nextel merged in 2005.)

Four Value Blocks

The arguments against selling the long distance wireline business were generally tied to four value blocks:

1. Owner’s economics
2. Business sales
3. Partner support
4. Difficulty of separation

The first value block argued that Sprint’s wireless business required owner’s economics for long distance transport. Wireless calls and data are generally only wireless from your handset to the nearest cell tower. Cell towers within a given area (e.g. a metropolitan area) are connected via copper and fiber cables to a regional hub and then calls and data are connected to their end destinations via national and global fiber optic networks. Because Sprint owned its own national and global fiber optic network, it was believed that the company benefitted from a lower cost than if it had to pay to use someone else’s fiber network.

The second value block argued that providing traditional long distance wireline voice and data services was necessary in order to sell wireless services to businesses. Sprint had good relationships with telecom buyers at most medium to large businesses because we had been selling them long distance voice and data services for decades. It was believed that these relationships were key if the company also wanted to sell those businesses wireless voice and data services.

The third value block argued that Sprint’s long distance fiber network was a critical element of our relationship with key partners. At the time, our most important partners were cable operators who owned and operated local fiber networks but generally didn’t own long-haul fiber to connect their local networks together. It was believed that providing low cost transport services using Sprint’s fiber was an essential element of our contribution to the partnership.

The fourth value block argued that Sprint’s wireless and wireline assets and operations were so tightly integrated that it would be very risky and expensive to try to separate them.

Cognitive Biases

Daniel Kahneman and Amos Tversky introduced the idea of cognitive biases and their impact on decision making in 1974. Kahmeman continued to develop the scientific understanding of biases and in 2002 was awarded the Nobel Prize for his work.

Some of the most common biases that impact business decisions include:

• Endowment bias
• Status-quo bias
• Sunk-cost bias
• Loss aversion bias
• Confirmation bias
• Self-interested bias

Endowment bias describes the perception that something we own is worth more than something we don’t own. More precisely, for a given asset, those that own it value it more highly than those that don’t own it. In Sprint’s case, the company probably too highly valued owning its fiber network.

The status-quo bias describes our natural comfort with things staying the same. The familiar saying “if it ain’t broke, don’t fix it” is a reflection of this bias (and actually often wise counsel). Both at a corporate level and on an individual level, I believe many key decision makers were comfortable with how the company transported traffic over its own fiber network and saw too much risk in changing to transporting traffic over someone else’s fiber network.

Sunk-cost bias shows up when businesses (or people) continue to invest money and effort into something because they don’t want to lose the value of the investments they’ve made in the past. Instead, it is usually best to evaluate the best use of your next dollar (or hour of work) independent of the decisions you’ve made in the past. Sprint was proud of the fiber optic network it had built and would rather keep investing in it than handing it over to a buyer and redirecting its investments to more value creating opportunities.

Loss aversion bias reflects our natural over-valuing of potential losses compared to potential gains. We are more worried about what we might lose than we are about missing out on what we might gain. We’d rather not lose $1 than gain $1. So, we do everything we can to avoid those losses, even if it means missing out on potential gains. There were some at Sprint who argued that outsourcing transport might actually prove to be financially worse for the company than continuing to operate our own fiber network and weren’t willing to risk that loss even if outsourcing had the potential to deliver great gains.

Confirmation bias shows up when we pay attention to information that supports what we believe and ignore or dismiss information that contradicts what we believe. I think we are all guilty of confirmation bias at different times, and I’m sure that it showed up in Sprint’s decisions about wireline.

Self-interested bias happens when people (consciously or sub-consciously) favor options that are in their self-interest. For many involved in Sprint’s wireline decisions, their identity and power were very much tied up in the company’s wireline network assets and long distance business. Selling the assets/business would likely negatively impact them and the people that worked for them. This undoubtedly impacted Sprint’s wireline decisions.

Value Block Realities

As I mentioned above, there were four value blocks used to argue against exiting the long distance wireline business. Biases may have kept us from universally accepting the realities around each of these value blocks, but let’s examine each one in turn.

Owner’s Economics
As a company, we believed that we needed to own our own fiber network in order to have a competitive cost basis in the very competitive wireless industry, but what was the truth? Was it really less expensive to own and operate a network than it was to buy capacity on someone else’s network? My team tried hard to find out.

Getting price quotes from a potential partner was easy enough (although getting the demand forecasts was harder than I expected). The challenge was getting the equivalent costs for continuing to own and operate the network. We could easily get incremental costs, but it seemed to be impossible to get reliable fully-loaded costs. In part this was tied to value block #4 — that separating the wireline and wireless networks would be hard. Getting true internal costs was challenged because it seemed impossible to separate out which elements of cost should be attributed to the wireline transport network and which to the wireless service network.

The most reliable indicator came from comparing Sprint’s cost position to those of peers who did not own their own fiber networks and were having to buy transport from other providers. Two notable examples were Nextel and T-Mobile USA. Both were wireless-only and both had higher profit margins than Sprint. Nextel also had higher pricing, which could help explain higher margins, but T-Mobile was the value player, offering discounted pricing relative to the industry. Since T-Mobile had higher margins than Sprint, it was hard to claim that Sprint needed to continue to operate it’s own fiber network. (But many continued to vehemently argue that need, or at least feared the possibility that exiting wireline would drive up costs.)

Another concern related to this value block was that we would end up having to buy fiber transport capacity from Verizon and/or AT&T, our two main wireless rivals. In the early 2000’s, many of the long haul fiber networks went bankrupt. Some were able to successfully exit bankruptcy, but the industry consolidated from about a dozen providers down to just a handful. As describe more thoroughly below, Verizon acquired MCI Worldcom, the #2 long distance provider and SBC acquired AT&T, the #1 provider, and renamed itself AT&T. It was feared that, in time, Verizon and AT&T would buy up the remaining long haul fiber networks enabling them to raise wholesale prices to Sprint. Especially if Sprint exited wireline, it is unlikely regulators would ever let this happen, but it was a legitimate concern.

Business Sales
When I joined Sprint in 2003, I helped develop the strategy for the division focused on selling to business customers. At the time, the company uniquely had wireless and long distance wireline capabilities and that became the core of our strategy. We believed that our portfolio of capabilities would enable us to deliver valuable solutions to business customers. (You can read more about how we developed that strategy here.) We thought that our long-standing relationships with the telecom buyers (typically part of the IT organization) at medium and large businesses would provide the natural entry for selling wireless services and integrated wireless/wireline solutions as wireless became increasingly important to businesses.

Our theory seemed to be validated by the actions of our competitors.

When the U.S. government broke apart AT&T, the long distance division retained the AT&T name. The local telephone business was broken into seven Regional Bell Operating Companies (RBOCs): NYNEX, Bell Atlantic, BellSouth, Southwestern Bell, Ameritech, USWest, and Pacific Telesis. AT&T originally did not see the growth potential of wireless so the nascent wireless operations went to the RBOCs. In the early 1990s, the company recognized its mistake and acquired McCaw Cellular/CellularOne. By 2000, AT&T had become an unwieldy conglomerate and began spinning off businesses. AT&T Wireless was spun off as an independent company in July 2001.

Verizon was formed through the merger of three large local telephone companies: Bell Atlantic, NYNEX, and GTE. The company strengthened its wireless capabilities in the late 1990s by rolling up the wireless operations originally in those three local providers plus Pacific Telesis and USWest. It continued to grow wireless through acquisitions throughout the 2000s, with the largest being Alltel in 2008. In 2005 Verizon acquired MCI, the second largest long distance wireline provider. At the time the company described the combination this way: “The merger will enable Verizon to better compete for and serve large-business and government customers with a full range of services, including wireless and sophisticated Internet protocol-based [wireline] services.”

Like its RBOC sister Verizon, Southwestern Bell spent much of the 1990s and 2000s rolling up the telecom industry. It acquired Pacific Telesis (but not its wireless operations), Ameritech, Southern New England Telephone, and BellSouth and renamed itself SBC. Southwestern Bell and BellSouth had combined their wireless operations into a joint venture called Cingular, and Cingular had acquired AT&T Wireless to become the largest wireless provider in the U.S. In 2005, SBC acquired long distance leader AT&T and rebranded everything AT&T. The company became the market leader in both long distance wireline and wireless services, and one of the two dominant players in local telephone services.

So, it seemed that the industry was following Sprint’s lead — pulling all the pieces together to offer a complete package of services and solutions to business customers. We must be right, right?

Unfortunately, having a multi-year lead in offering both wireless and wireline to business customers didn’t translate into market success. Sprint had great success in consumer wireless, but trailed significantly in business wireless sales.

In 2004, my team took on the task of developing a plan for becoming the leader in business wireless solutions. We performed extensive market research and built sophisticated models. In the end, we learned that there were two viable paths to business wireless solution leadership. One path was the one Sprint, Verizon, and AT&T had been pursuing — leverage existing telecom/IT relationships to introduce wireless solutions. The second path was to circumvent IT and sell directly to business leaders with distributed and mobile workforces. We recommended the second path.

Before we completed our work, Sprint announced its merger with Nextel. Nextel was the clear leader in business wireless and had been pursuing the second path. In fact, Sprint had made a big splash when it announced a strategic partnership with Ford Motor Company, that included Sprint becoming the official wireless provider to the car maker, but after the merger we found that Nextel had sold significantly more wireless handsets to Ford than Sprint. They had done so by bypassing IT and going directly to Ford dealerships around the country. Nextel sold the benefits of increased productivity to buyers who quickly saw the return on investment. In contrast, Sprint went through lengthy purchasing processes with IT leaders who were focused on reducing costs and protecting against security risks. Nextel sold to those focused on capturing the power of the potential of wireless, while Sprint sold to those focused on managing the danger of wireless. Nextel sold more, faster, and at a price about 25% higher than Sprint’s. Did I mention that Nextel did not have a wireline network?

Partner Support
As I mentioned above, in 2004 Sprint’s most important partners were cable operators. We believed that providing these partners with transport services across the Sprint wireline network at a very aggressive price (near our own cost) would be seen as a valuable contribution to the partner relationship. In reality, the cable companies generally didn’t rely on Sprint for their long distance transport needs. They were able to buy what they needed on the open market at prices competitive with what we offered.

Separation Challenges
A telecom point of presence (POP) is typically laid out in rows of equipment. Each row is made up of bay after bay of equipment. Each bay has multiple shelves, and on each shelf are slots into which cards slide. The cards are dense circuit boards with a variety of electronic components that combine to perform a specific function for the network. If you’ve ever been in a library, you can picture a telecom POP by imagining each bay being a separate bookcase, with multiple shelves, and the books on the shelf being equivalent to the electronic cards.

Photo by Manuel Geissinger

As Sprint built out its network, it didn’t build separate wireline and wireless networks. It put wireless and wireline equipment in the same rows and sometimes the same bays. Sometimes a given shelf might have cards that served both wireless and wireline needs. Wireless equipment was directly cabled into the wireline network to drive performance and efficiency. Separating the two networks would be hard, but not impossible.

What Did Sprint Do in 2004?

As you’ve probably guessed, in 2004 Sprint did not decide to exit the wireline business, but rather decided to prune it. We eliminated products that weren’t aligned with future growth and with the vision for integrating with wireless. Some of these revenue streams we sold to competitors to generate cash to invest elsewhere. We focused on markets (geographic and vertical) where we were best positioned to deliver our integrated wireless and wireline promise. We shut down sales offices and let people go. We also invested in the future product portfolio we envisioned. We believed this could extend the financial performance enough to make wireline worth keeping well beyond the original 3 year forecast.

Apparently the company decided to keep wireline for another 18 years, although analysts believe that it has been unprofitable for a few years. But maybe our decision wasn’t such a bad one. I still think we might have been better served to find a buyer for wireline who could invest to provide us with even better capabilities, with greater scale enabling even lower costs, and who would pay us a fair price for the assets and customer contracts that we could reinvest in growing our wireless business.

But then again, I’m sure my thinking is biased in some way in believing that would have been a better outcome.

What Did T-Mobile Do in 2022?

T-Mobile sold what had been the Sprint long distance wireline business to Cogent Communications for $1. The company also agreed to buy transport services from Cogent, committing to spend at least $350 million in the first 12 months and another $350 million over the next 3.5 years. T-Mobile will also help cover the separation costs for employees laid off as a result of the transaction. As Cogent works to integrate what they’ve acquired, and to further prune the business to make it profitable, hard decisions will undoubtedly be required. The T-Mobile/Sprint wireline business currently has around 1,300 employees, so sadly there may be many good people losing their jobs.

Although the $1 sale price may sound like a bad deal for T-Mobile, especially in light of the millions T-Mobile is committing to pay Cogent, I think the company is wise to finally make this decision. T-Mobile’s outsider perspective allowed it to avoid the cognitive biases that have kept Sprint from making that decision for nearly two decades.

Well done T-Mobile!