Connectivity Provider Margin Compression Since 1980

Margin compression is a problem of long standing in the connectivity industry. Generally speaking, profit margins have declined over time, partly because of new competition, partly because of better technology, partly because of government rules promoting lower prices. 

Generally speaking, profit margins were highest in the monopoly era, when telecom companies operated as sanctioned monopolies. As newer markets have operated in a competitive manner, profit margins have been lower, even if mass market adoption has been helped by those lower prices. 

Product	Years	Profit Margin (%)
Fixed-line Voice	1860-1980	50-70
Mobile Voice	1980-2000	30-50
Internet Access	2000-2010	20-30
Content Services	2010-present	10-20

Calling rate trends clearly show the impact of competition and better technology (including voice over IP calling and messaging substitutes; substitution of email and texting and messaging for voice communications). 

Voice Calling Price Changes (Per Minute) Since 1980
Date	Product	Rate (in USD)
1980	Local	0.25/min
1985	Local	0.20/min
1990	Local	0.15/min
1995	Local	0.10/min
2000	Local	0.05/min
2005	Local	0.02/min
2010	Local	0.01/min
2015	Local	0.005/min
2020	Local	0.002/min
1980	International	$1.00/min
1985	International	0.80/min
1990	International	0.60/min
1995	International	0.40/min
2000	International	0.20/min
2005	International	0.10/min
2010	International	0.05/min
2015	International	0.02/min
2020	International	0.01/min
1980	VoIP	N/A
1985	VoIP	N/A
1990	VoIP	N/A
1995	VoIP	N/A
2000	VoIP	$0.05/min
2005	VoIP	$0.02/min
2010	VoIP	$0.01/min
2015	VoIP	$0.005/min
2020	VoIP	$0.002/min

One can see the same trend for home broadband or wide area network data transport prices. 

To the extent that WAN transport revenue has not plummeted directly, it is because volumes of data to be moved have increased so much. 

But there have been other effects, such as the displacement of most traditional telcos as meaningful providers of WAN data transport, compared to private networks operated by hyperscalers and other long-haul data transport specialists.

Year	Data Transported (in petabytes)	Cost per megabyte (in USD)
1980	0.01	$100,000
1985	0.1	$10,000
1990	1	$1,000
1995	10	$100
2000	100	$10
2005	1,000	$1
2010	10,000	$0.1
2015	100,000	$0.01
2020	1,000,000	$0.001

"What Comes Next?" Will Eventually Matter

Whatever comes next as the key revenue driver and source of profit for connectivity providers, the key product will necessarily have to be a “communications” product of some sort. Since 1860, when “telephone service” was invented and commercialized, the key driver of industry revenue and profits has been some connectivity service.

Between 1860 and 1920 the primary revenue driver was the ability to place a local phone call, though subscriptions were not likely greater than about 10 million by the end of the period.

Starting in the late 1920s the ability to make international calls became a major revenue driver and the driver of industry profits. That continued to be the case until the late 1970s. But competition in long distance disrupted profits.

After the early 1980s, mobile phone service emerged as the driver of industry revenue growth, and finally the dominant revenue source and profit driver for the global industry. 

And while mobility services now drive a clear majority of all revenues, on fixed networks there has been an evolution from voice calling to internet access as the key revenue item. 

“What comes next?” is a logical question, as hard as it might be to envision a time when mobile phone services were not the industry revenue driver or its key source of profits. As computing eras have given rise to new products and value drivers, the connectivity business has moved through eras where the dominant revenue model has shifted. 

Right now I am hard pressed to envision what core connectivity service emerges to displace mobility as the industry revenue and profit driver. But replacements eventually will be developed. 

That evolution has happened in the connectivity business and the computing business as well. 

Computing cycles used to happen only on mainframes or minicomputers. Then we shifted to personal computers, then to client-server models, then to smartphones and embedded instances (calculators, cash registers, elevators and an ever-wider range of consumer electronics. 

Well over half of all computing cycles now happen on phones, according to Statista and IDC estimates. 

What comes next as an era of commuting is not so easy to categorize, though a new computing era is sure to arrive. One way some of us might point to what is coming is to argue that “the next 10,000 startups will essentially be based on “take X and add artificial intelligence.” 

Some might argue we now are in the era of cloud computing. Others will prefer the term “mobile computing.” Terms such as “pervasive computing” or “ambient computing” have been floated as well. 

Whatever the preferred term, computing now is highly distributed. Some might note that up to 46 percent of computing cycles happen on cloud computing fabrics. 

In both computing and connectivity industries, a next era will develop. The logical path seems clearer for computing, though, where applied artificial intelligence seems to underpin every major trend and development: metaverse, augmented reality, search, content consumption and creation, robotics, automation, advertising and shopping and education. 

Beyond the ability to make a phone call, international calling, mobile phones and internet access have been the key shifts in the connectivity business since 1860. As hard as it might be to contemplate, something else eventually is going to emerge. 

Edge computing, private networks, fixed wireless and internet of things might help. But none of them presently seems sufficient to achieve the mass scale to displace mobile phone service as the industry revenue and profit driver.

Will Investable Categories of Digital Infrastructure Broaden?

It is perhaps predictable that, at some point, private equity investors will start to broaden the categories of assets they seek to acquire. Some of the largest firms have already begun to invest in some assets beyond data centers, cell tower networks or optical fiber access networks.

Giant Blackstone has made some investments in software assets including artificial intelligence and cybersecurity, some might note. KKR likewise seems to have made some early steps to broaden beyond data centers and fiber networks, and into software. 

Apollo Global Management likewise is making similar moves, as is Macquarie Infrastructure and Real Assets. The issue for many smaller funds and firms is that the expertise to conduct prospecting and due diligence in the newer areas requires expertise not presently onboard. 

It also will be harder to find assets that offer the same predictable cash flow as the physical assets. But assets such as Intrado, which supports emergency calling capabilities, and was recently acquired by Stonepeak, provide an illustration. 

Still, some investments arguably are more risky, from a predictable cash flow perspective. 

Apollo Global Management, for example, recently took Intrado private, for example. Intrado is perhaps best known for supplying systems supporting emergency calling features and services. So the issue is how many more such franchises might be available in spaces adjacent to classic digital infra. 

At first glance, that might be somewhat rare, though over time some quasi-franchises could develop in specialized operating systems, perhaps database management and likely security. Some might propose virtualization software used by connectivity providers as a possible future opportunity, once marketplace standardization happens. 

Others might propose that dedicated internet access networks merit consideration, especially when provided by specialist firms. Content delivery networks are another possibility. And though most of the funding for artificial intelligence firms will come from venture capitalists, eventually some of those assets will mature to the point where PE gets interested.

Hedge funds arguably are more likely to take stakes in software or hardware firms than private equity. Elliott Management, properly a hedge fund, acquired Gigamon, a supplier of security products.

Firms that specialize in software deals also might be buyers of assets that could overlap with digital infra investors at some point. Coupa was acquired by Thoma Bravo, which specializes in software deals. Arguably most software investments are made by venture capital firms, at least for startups. 

As a rule, the attraction includes the expectation that financial performance, and therefore asset value, can be enhanced by the private equity owners and managers. In some cases, perhaps nearly all cases, digital infra assets (which are a mix of data center, cell tower and access network assets) are valued more highly than operating businesses using their own infrastructure. 

Also, firms that use platform business models (outside the connectivity businesses) also tend to have higher valuations than connectivity operating businesses. Since true platform business models are rare in the connectivity business, and tend, even when present, to represent only a smallish portion of total revenues, it remains unclear how valuations could develop. 

But it might be reasonable to expect a further boost in valuation metrics of some magnitude. 

Valuation Metric	Digital Infrastructure	Telco	Platform Business Models
Price-to-earnings (P/E) ratio	20-25x	15-20x	30-35x
Price-to-sales (P/S) ratio	10-15x	8-12x	15-20x
Price-to-book (P/B) ratio	1-2x	0.8-1.2x	1.5-2x
Enterprise value (EV)/EBITDA	20-25x	18-22x	25-30x
Dividend yield	3-5%	2-4%	1-2%

Obviously, any such comparisons are suggestive, as ratios change based on other conditions such as the size of the firm, the geographies where it operates, firm growth rates, product mix and degree of competition in its markets. 

That is why managed service firms in the connectivity industry, deemed to operate with higher or specialized value added, tend to earn higher valuation multiples, all other things being equal. 

Valuation Metrics	Telco	Cable TV Operator	Managed Service Provider Businesses
Price-to-Earnings Ratio (P/E Ratio)	20.25	17.93	28.10
Price-to-Sales Ratio (P/S Ratio)	1.65	2.29	8.31
Enterprise Value-to-Revenue (EV/R) Ratio	1.73	2.56	18.07