Are Telcos "Trapped" by Language or Operating Metrics?

Sebastian Barros, Circles managing director, might be quite right that average revenue per user and gross additions (new subscribers) no longer make sense. The “KPIs (key performance indicators) they still cling to are built for a world of selling minutes and megabytes, not for today’s digital ecosystems,” he argues.

One might agree with the sentiment, but the logic misses something important. It might be correct to argue that “consumers no longer just buy ‘minutes and megabytes,’” as Barros somewhat rightly notes. “They buy ecosystems: streaming, wallets, commerce, and cloud.” 

Perhaps more correctly, consumers use ecosystems, wallets, streaming, commerce, cloud, minutes and megabytes. What we actually buy from mobile and fixed network access providers is the use of their network resources so we can talk, text and use the internet. 

But “telcos” and “internet service providers” are not in businesses where metrics such as units shipped, book-to-bill ratio or monthly active users or advertising revenue per user make sense as operating indices that inform us about how the business is faring. 

Where ISPs sit in the value chain means they make their money selling internet access and some communication services. So ARPU, churn rates and net additions actually are the relevant operating metrics. 

Those metrics will not make sense for others in the value chain. But ARPU, gross adds, net adds and churn are very much the right numbers to track operating performance. 

Here’s the point: ARPU and gross adds do not “trap” ISPs in old ways of thinking. Those metrics are the right ones for the business they are in and where they are situated in the value chain. ARPU and gross adds might make no sense for other participants in the value chain. 

But just because “ARPU and gross adds make no sense for the business I am in” does not mean “they make no sense for the business ISPs are in.” 

Segment	Example Players	Key Operating Metrics	Typical Characteristics
Semiconductors / Chip Suppliers	Nvidia, Intel, AMD, Qualcomm, Broadcom	- Gross Margin (% of sales) - Fab Utilization Rate - R&D Intensity (% of revenue) - Average Selling Price (ASP) - Units Shipped	Capital-intensive; cyclical demand; high fixed costs; margins vary by product (leading-edge chips much higher).
Networking / Hardware Equipment	Cisco, Ericsson, Huawei, Arista	- Gross Margin - Book-to-Bill Ratio - Installed Base Growth - CapEx as % of Revenue - Service Revenues %	Dependent on telco/enterprise capex; recurring service revenues important; margin pressure from competition.
Access Providers (Telcos, ISPs)	AT&T, Verizon, Comcast, Deutsche Telekom	- ARPU (Average Revenue per User) - Churn Rate - Network CapEx (% of revenue) - EBITDA Margin - Subscriber Growth	Capital-intensive, slow growth, regulated; sticky customer base but high infrastructure costs; moderate margins.
Cloud / Hyperscalers	AWS, Microsoft Azure, Google Cloud	- Revenue Growth Rate - Gross Margin - Utilization of Datacenters - Customer Retention - Operating Income Margin	High scalability; strong growth; capex heavy but high-margin once scaled; sticky enterprise contracts.
Search Engines	Google, Bing, Baidu	- MAUs (Monthly Active Users) - ARPU (via advertising) - Ad Load (ads per page/search) - Click-through Rate - Operating Margin	Network effects; ad-driven monetization; extremely high margins; market concentration.
Social Media Platforms	Meta, TikTok, Snapchat, X (Twitter)	- DAUs / MAUs - ARPU (ad revenue per user) - Engagement Time per User - Ad Fill Rate - Operating Margin	Strong network effects; monetization via ads; margins high, but sensitive to user growth and ad demand.

It is quite true that “pipes and bundles are the only game in town,” as Barros notes. Unfortunately, most of those other businesses in the value chain are not the ones telcos and ISPs are in, for the most part. 

I find it is common in business for people to make a certain sort of mistake. They might argue something “cannot be done” because “my company, with its resources and business model, cannot do so.” That never means some other company, with different resources, cannot succeed in doing so.

Nor, in this case, do the prevailing operating metrics “trap” participants. Telcos and ISPs are access providers. It is their role in the value chain. If they were to substantially change roles, and move elsewhere in the value chain, of course the operating KPIs would change. 

But ISPs are not limited by their choice of metrics: they are “limited” by their chosen roles in the value chain.

What is the Value of a Copper Access Line That Cannot be Upgraded to Optical Fiber?

Some of us used to wonder what many telcos were going to do as they phased out their copper access facilities, since many are still covered by older laws mandating they provide service to any customer who asks for it. The problem is that those telcos must still have the means to provide service, even if they cannot use copper or optical fiber facilities. 

Wireless, in some form, always has been the only realistic alternative. Whether that is satellite service, mobile phone service or fixed wireless, untethered platforms might always be the only way to provide universal service in areas where a fixed network using cables is impractical. 

The latest advances allow standard mobile phones to communicate with low earth orbit satellites without any extra gear or software. 

Which still leaves us with the problem of how to value stranded copper assets, which are declining in value, especially when they cannot be upgraded to optical fiber, and as consumers continue to migrate away from use of fixed networks for voice services. 

Many potential fixed network “home broadband” assets owned by smaller telcos must use a blended valuation approach, as such firms generally own a mix of copper access lines that cannot be profitably upgraded to fiber-to-home; copper lines that can be upgraded as well as existing fiber lines that are operational.

Each asset has distinct valuation characteristics, with built fiber lines being the most valuable, non-upgradable copper lines the lowest valuation. 

The big U.S. telcos arguably always have grown more from acquisitions than from organic growth. Verizon, AT&T, Charter, Comcast and T-Mobile, for example, have done so. In the FTTH business, which always is capital intensive, there also is a “time to market” advantage. 

Building lines takes time. And even when built, take rates might remain in the 40-percent range. In other words, up to 60 percent of the assets are essentially stranded, with no revenue produced. So buying lines that do produce revenue (have subscribers on them) has lots of value. 

If building FTTH lines costs between $800 and $1200 per location, and we assume 40-percent take rates, the cost per subscriber can range from $2000 to $3000. 

If we add in marketing and acquisition costs, the full cost of provisioning a line with a customer on it can range from $2350 to $3700, assuming sales and marketing cost between $200 and $400 per sub, plus installation costs between $150 and $300 per subscriber. 

Scenario	Cost per Passing	Take Rate	Cost per Subscriber
Low Cost	$800	40%	$2,000
High Cost	$1,200	40%	$3,000
Average	$1,000	40%	$2,500

If that is the case, then it often makes sense to pay a premium to acquire customers and facilities rather than build them, if “buying rather than building” accelerates cash flow and also allows some additional economies of scale. 

When copper access lines are upgradable, there is additional upside. And even dwindling copper access revenues provide some amount of revenue and cash flow, even if they are declining assets. For providers who own mobile network assets, there additionally is the potential to serve former copper customers with fixed wireless access as well, as a longer-term alternative to using copper access. 

Eventually, that also will change the valuation of access network assets.

Bye Bye Skype

As Microsoft retires Skype in favor of Teams, it might be useful to recall just how impactful “voice over IP” services such as Skype were in dismantling the telco profit engine.

For example, looking only at revenue, in 2000 global international call revenues were in the range of $80 billion to $100 billion, with very-high profit margins. By 2020, international calling revenues had dropped to about $15 billion to $20 billion, with profit margins compressed. 

 

Decline in International Calling Revenue with VoIP Adoption (2000–2020)
Year	Est. Int'l Calling Revenue (Billion USD)	VoIP Adoption & Skype Milestones	Notes
2000	~80–100	Minimal VoIP presence; traditional PSTN dominates	High tariffs for international calls; telecom monopolies prevalent.
2003	~75–90	Skype launched; 11M users by 2004	Skype introduces free VoIP calls and low-cost PSTN calls, challenging telecom pricing.
2005	~70–85	Skype acquired by eBay ($2.6B); 54M users	VoIP gains traction; telecoms begin lowering rates to compete.
2008	~60–75	Skype grows to 405M users	Economic recession impacts telecom revenue; VoIP alternatives expand (Viber, WhatsApp emerging).
2010	~50–65	Skype disables third-party integrations; 663M users	Telecoms lose market share to VoIP; mobile data plans begin reducing VoIP dependency.
2013	~40–55	Skype-to-Skype int’l traffic up 36% (214B minutes)	TeleGeography notes VoIP capturing significant call volume; traditional revenue continues to decline.
2015	~30–45	WhatsApp, Viber, and others compete with Skype	Mobile apps erode Skype’s dominance; telecoms shift to data-driven models.
2018	~20–35	Skype daily users at 40M (2020 peak)	VoIP services saturate the market; telecom firms  focus on broadband and mobile data revenue.
2020	~15–25	Skype usage spikes 70% during COVID-19	Despite Skype’s decline to 36M daily users by 2023, VoIP remains dominant; traditional int’l calling revenue nears obsolescence.

Profit margins were an important part of the early, pre-VoIP story. Net profit margins on international voice were as high as 25 percent back in 2000. Current net margins are in the range of three percent to possibly five percent. 

In a real sense, VoIP services including Skype disrupted the telecom industry profit driver. 

Year	Domestic Long-Distance Margin (%)	International Long-Distance Margin (%)	Discussion
2000	~12–18%	~15–25%	High margins due to limited competition and high per-minute rates. Domestic margins are slightly lower than international due to local competition. Estimated from telecom sector data and peak long-distance revenue.
2001	~12–17%	~15–24%	Stable margins but early pressure from mobile and VoIP adoption. International margins are higher due to termination fees. Estimated from sector trends.
2002	~11–16%	~14–23%	Decline in domestic voice revenue began as mobile plans offered "bucket" minutes. International margins remained higher but faced VoIP competition. Estimated from sector data.
2003	~10–15%	~13–22%	Continued erosion from mobile and VoIP (e.g., Skype). International margins supported by high termination rates. Estimated from sector trends.
2004	~9–14%	~12–20%	VoIP and internet-based calling reduced costs and rates, squeezing margins. Domestic margins lower due to flat-rate plans. Estimated from sector data.
2005	~8–13%	~11–18%	Long-distance business peaked in 2000; by 2005, revenues were declining rapidly. International margins are higher due to mobile international calling demand. Estimated from sector data.
2006	~7–12%	~10–17%	Domestic margins hit by unlimited calling plans and VoIP. International margins supported by slower price erosion in mobile long-distance. Estimated from sector trends.
2007	~6–11%	~9–16%	Domestic long-distance became commoditized; international margins pressured by OTT apps (e.g., WhatsApp). Estimated from sector data.
2008	~5–10%	~8–15%	Long-distance revenues halved from 2000 peak. Economic recession and VoIP adoption further reduced margins. Estimated from sector data.
2009	~5–9%	~8–14%	Smartphone adoption and VoIP apps (e.g., Skype) eroded margins. International mobile long-distance retained higher margins. Estimated from sector trends.
2010	~4–8%	~7–13%	Domestic long-distance margins near sector average (~4.82%). International margins are higher due to termination fees and mobile demand. Estimated from sector data.
2011	~4–7%	~7–12%	Domestic margins low as carriers bundled unlimited long-distance. International margins declined due to VoIP growth. Estimated from sector trends.
2012	~3–7%	~6–11%	Domestic long-distance fully commoditized; international margins affected by outsourcing and fraud. Estimated from sector data.
2013	~3–6%	~6–10%	Mobile data surpassed voice revenue; domestic long-distance margins were minimal. International margins supported by the wholesale voice market. Estimated from sector trends.
2014	~3–6%	~5–10%	Voice over LTE and VoIP reduced standalone voice profitability. International margins pressured by low-cost VoIP providers. Estimated from sector data.
2015	~3–5%	~5–9%	Domestic long-distance margins were negligible as unlimited plans dominated. International margins declined due to OTT apps. Estimated from sector trends.
2016	~3–5%	~4–8%	Voice services bundled with data; domestic margins near zero. International margins are low but supported by wholesale carriers.
2017	~2–5%	~4–8%	Domestic long-distance margins minimal; international margins affected by grey routes and fraud. Estimated from sector trends.
2018	~2–4%	~4–7%	Domestic margins near zero as voice bundled with data plans. International margins low due to VoIP and 5G adoption. Estimated from sector data.
2019	~2–4%	~3–7%	Voice commoditization is complete; international margins slightly higher due to wholesale voice demand. Estimated from sector trends.
2020	~2–4%	~3–6%	COVID-19 increased communication demand, but voice margins remained low due to free VoIP apps. Estimated from sector data.
2021	~2–4%	~3–6%	Domestic long-distance margins negligible; international margins low but supported by enterprise demand. Estimated from sector trends.
2022	~2–4%	~3–6%	Telecom services margin ~4.82%; domestic voice margins near zero. International margins are low due to wholesale price wars. Estimated from sector data.
2023	~2–4%	~3–5%	North America wholesale voice market faced intense competition, eroding international margins. Domestic margins are negligible. Estimated from sector trends.
2024	~2–4%	~3–5%	Wholesale voice market valued at $40.26 billion in 2025, but margins low due to VoIP and 5G. Domestic margins near zero. Estimated from sector data.

Also, VoIP was not the only huge driver of a shift in consumer behavior. “Calling” became something most people did on their mobile phones. 

Fixed-network revenue dropped from $200 billion globally in 2000 to under $50 billion by 2020, while mobile revenue grew from $500 billion to $1.6 trillion, for example. U.S. telco revenues likewise shifted from fixed to mobile; legacy voice to VoIP. 

U.S. Telco Revenues 2000 to 2024
Year	Mobile Voice	PSTN Voice	VoIP
2000	10	100	1
2010	60	60	21
2020	110	20	41
2024	130	4	49