Thursday, October 8, 2015

AT&T Voice over Wi-Fi: Feature, Not a Service

A waiver granted by the U.S. Federal Communications Commission to AT&T now allows the firm to offer Wi-Fi calling. The business model is not yet fully visible, as the service has not yet been fully launched.

The context is that Apple’s iPhone, since iOS 8, has supported voice calling using Wi-Fi connections. But full value also requires that the service interwork seamlessly with carrier voice. Until now, that has not been possible.

The new feature illustrates the challenge of voice business models. At least so far, voice calling using Wi-Fi is mostly a capability, not a direct revenue driver for AT&T.

As with use of Wi-Fi for Internet access, the feature might be most useful, in the U.S. mobile market, for callers in areas where indoor mobile signal is weak and Wi-Fi signals are strong.

Though some tier-one service providers have launched their own voice over IP services, they arguably have gained little traction, compared to the third party app and service providers.

In a broad sense, the notion that access providers could compete successfully with over the top providers in voice has proven incorrect.

VoIP has mostly shrunk the retail revenue opportunity for voice, and shifted demand to third parties, and away from carriers.

Eventually, the feature might have greater indirect revenue implications, however.

As one or more cable TV operators enter the mobile market, they are expected to lean on their own hotspot networks and Wi-Fi for network infrastructure.

In that case, voice over Wi-Fi will help the overall business model, offloading demand from mobile to the fixed network.

That could have direct financial implications. To the extent that cable TV companies rely on wholesale access provided by other mobile operators, offload to Wi-Fi will mean lower payments to the wholesale services provider.

No comments:

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...