The whole issue of telco profits is complicated, but one potential path to addressing the issue is to separate infrastructure from services—what’s often described as a netco/servco split. The notion is discussed in a Tata white paper, but I want to develop, justify, and offer my own views here.
This turns out to be a difficult topic to get direct operator input on, because there’s a sense of confidentiality about, both among operators who have already made the move and among those who might be considering it. I got most of my comments (93 people, representing 61 operators) from either retired people or lower-level planners, and no CxOs had anything to say.
Right from the start, there’s an irony here. Everyone agrees that the idea of the split emerged out of the original “privatization” push of the 1980s. Then, the goal was to level the playing field between national incumbents and “competitive local exchange carriers” or CLECs. To prevent incumbents from leveraging their position, they were required to offer any higher-level services (in the US, regulators called these “information services”) through a fully separate subsidiary. The thing is, the presumption was that the combined company would have more competitive strength, which means this early example of splitting actually argues against it offering any benefit.
One clear negative to a split is that it eliminates subsidization of low-level services from new higher-level stuff, which is bad. Another is that it probably makes it harder to build new infrastructure like edge computing, by reducing the ROI to be earned. There has to be compensatory positives here, or splitting is dumb.
The most-cited of the non-cynical arguments for the split are that it would break the telco-think inertia we see so much of these days, the fixation on the past glory of connection services and field-of-dreams planning. My problem with this is that you could achieve the same thing by giving a “service division” autonomy, and everyone agrees that doing the split is a major undertaking.
Another theory is that it would open the netco to create service partnerships with others, like cloud providers. Of course, this would mean encouraging retail competitors to the newly separated servco, and it would benefit netco only if those competitors were better than the emergent servco. Then we have to wonder if this sort of thing wasn’t a deliberate form of what operators bemoaned for decades, OTT disintermediation. Still, we’ll get back to this one because it seems on the surface to be the best non-cynical option.
I don’t think that any non-cynical justification for splitting is a slam dunk, which leaves us to look at the cynical justifications. As it happens, both retired operators and current planners agree on the same one here, and so do I. What is it? That the split would make subsidization more palatable.
A netco with a wholesale obligation is less a monopolistic risk because its assets are contributing to many players. Big tech subsidies, or any industry-contributed subsidies, would go down easier if the subsidies were not paying a competitor. The essential notion would be to at least permit the separated netco’s return to a regulated monopoly status, which would make subsidization easier.
But are they necessarily inevitable? Remember the concept of the neutral host? Tower companies are netcos, or even a layer lower. But they work because the alternative is competitive overbuild. This brings us back to the second of our non-cynical justifications for a split. How much overbuild we could expect the servco level, and whether it could generate positive gains for a netco, depends on whether servco competition would generate new retail revenue. If it did not, then it could actually increase commoditization.
The problem I’ve already noted here is that it’s difficult to justify a claim that a servco is much different from an OTT. We have them now, because the Internet creates something that, in terms of features, isn’t much different from a sort of universal netco. Thus, it’s hard to see how servo competition could boost the retail service spend, which means there’s no new money to spend on netco services. Without new features there’s no new value, and without new netco capabilities any servco is just like today’s OTTs, so why would they spawn new features? This problem is fatal for the non-cynical justifications for the split, and maybe for all justifications.
That’s because the problem with the netco/servco split lies with the split itself, more than with how to justify it. If you do the split, I think it solidifies the notion of servco competition, a competition. Servcos, particularly big established ones like cloud providers, would wholesale netco features and add them to their own billing, to boost profits, not invest in new features at the network level. Thus, absent new enabling features from the netco, this will likely to lead not to more services but to commoditizing. And why would netcos invest? You have nothing to drive either netco revenue or netco investment, so you’d almost mandate subsidies.
Of course there’s no guarantee that subsidization or even nationalization will really work. Australia did something very like it with NBN, which has failed to gain full support or meet its original objectives. Subsidies are a public policy decision no matter who makes the direct payment, and public policy is fickle. Another certainty is that they don’t promote innovation; you ever hear of an innovative bureaucrat? That’s important because in the long run, innovation is essential to the success of network operators, and as long as it’s possible our industry has a future. Steps that stifle it are bad, and we’ve taken some already. Steps that kill chances for it simply cannot be allowed to happen.
My view here is simple. There are a lot of paths that lead to a stagnant networking industry, but most of them will offer us warning signs that give us time to take action, and most will keep the industry in a survivable state to almost the end. The netco/servco split isn’t one of these. I fear that if this step is taken, we risk a major problem with networks and online serviced that will prove difficult to fix. Policy-makers take note; this could be a very bad idea.
