Comcast’s announcement on its earnings call last week that it is triggering the MVNO relationship with Verizon and plans to test a ‘WiFi First’ wireless service over the next few months will once again shine a spotlight on the relationship between cellular and WiFi. My view is there are still some significant usability challenges that must be addressed before a ‘Wi-First’ service, as I call it, is ready for prime time.
Wi-First services along the lines of what Comcast plans to offer have already been available for a few years from MVNOs such as Republic Wireless and Scratch Wireless. As well, Intel and FreedomPoP recently announced they will be working together on a new phone for FreedomPop’s Wi-First service. Their services, available on specific devices, are positioned as a less expensive alternative to a typical cellular plan, offering unlimited voice, text, and data while on WiFi, with the option of using the cellular network as a backup when WiFi is not available. Google’s ‘Project Fi’ adds a twist of choosing the ‘always best connected’ scenario between WiFi and two cellular networks: Sprint and T-Mobile. Cablevision has been offering its Freewheel service for about a year, but the fact there is no cellular option when outside range of a hotspot is a major drawback.
So far, ‘Wi-First’ services have not taken off. The services are available only on select, purpose-built devices. They have not been heavily marketed. There have been some issues with the user experience. And cellular prices have been dropping. For those looking for a bargain, prepaid services such as Cricket and Straight Talk offer plans for as little as $30 per month.
A Comcast MVNO however, will up the Wi-First ante, by virtue of its size (23 million broadband subscribers, 55 million households passed, ~150 million ‘POPs’), and the fact it has deployed 10 million hotspots as part of the Cable WiFi consortium that also includes Time Warner, Cox, and Brighthouse. These Cable WiFi hotspots are a combination of about 9.5 million residential gateways (APs) equipped with a dual SSID (homespots) and 500,000 additional indoor and outdoor APs that Comcast and the other cable MSOs have deployed. The density of coverage in some neighborhoods is quite compelling. Comcast has marketed Xfinity WiFi as a convenient and useful broadband extension for its existing customers and as a way of saving on cellular data usage.
The idea is Comcast’s wireless service would leverage these millions of APs and would use Verizon’s wireless network when not in range of a WiFi hotspot. Presumably, Comcast’s service would be less expensive than conventional wireless plans and would be a convenient add-on for its existing subscribers, delivering Comcast potentially $30-40 in additional monthly ARPU for each customer that takes the service (before payments to Verizon).
Note also that, even if on the surface a Wi-First service might appear to be a threat to the mobile operators, they benefit from the MVNO relationship and the fact that data traffic is being offloaded from the cellular network.
The early Wi-First providers have done some pioneering work in developing a seamless experience between cellular and WiFi. This includes the ability for cellular features such as text messaging and voice mail to be delivered when the phone is in WiFi mode. The handoff of voice calls between cellular and WiFi works some, but not all, of the time. And WiFi voice quality can be excellent when the stars align between AP signal, device, software, and so on.
But, in my view, there are two major usability bugaboos that need to be addressed before Wi-First services are ready for the mass market. First, there continue to be issues related to how phones switch between ‘WiFi mode’ and ‘cellular mode’. Put simply, smartphones still often default to a WiFi hotspot even if the signal is weak or it is congested. This issue will become even more pronounced when voice is added to the equation, since voice quality really falls off a cliff when the signal is poor or the AP is congested. There also continue to be issues of having to re-authenticate to hotspots that are already known, which impacts the idea of a seamless Wi-First user experience.
In my use of Cable WiFi outside the home for example, I frequently encounter inexplicable “cannot join” messages, or experience very slow speeds, even where it appears there’s a good signal from the Cable WiFi AP. This also frequently happens in public locations that the phone ‘remembers’, such as at airports and coffee shops. In fact, in an increasing percentage of instances when away from the home or office, I find myself turning off WiFi because the cellular network is more consistent, reliable, and, ironically, often faster. Plus, cellular data costs less than it used to. This is a far cry from a few years ago, when customers invoked WiFi wherever they could.
Apple has clearly recognized this issue, offering a new feature called WiFi Assist, which switches the iPhone to “cellular” mode when the WiFi signal is weak. However well intentioned, Apple has received some backlash because it didn’t do a lot to inform customers of this feature, which can cause a spike in cellular data usage charges.
The second major issue is that Wi-First services must be available on the majority of devices users already own. Today, services from MVNOs such as Republic, Scratch, and Google only work on a handful of devices customized with purpose-built software for Wi-First. If Wi-First services aren’t available on the leading iOS and Android devices, they’ll never have broad appeal.
Fortunately, there’s progress on this issue. WiFi calling is now supported in iOS and is becoming more broadly supported by the operators in addition to T-Mobile. Voice over WiFi works pretty well, though there still continue to be issues when multiple users/devices contend for a hotspot and with handoff to cellular when leaving an AP. The spread of VoLTE should ease some of the handoff issues, since it’ll all eventually be data anyway.
There is a lot of complexity involved in delivering a reliable, seamless, and high quality service combining cellular and WiFi, which is why Google, Comcast, and others have been decidedly cautious in their approach. There is no panacea, just a lot of incremental work on both the network and device sides to tune and refine the experience. The major usability issues must be addressed before a “WiFi First” service is ready for the mainstream, mass market.
One thought on “Cellular and Wi-Fi Need to get Along Better”
Isn’t the major issue what happens after the signal reaches the access point, whether cellular or wifi ? Carriers own their infrastructure, and it is mostly optimized for voice (low latency), while the Internet isn’t equipped to reliably prioritize low-bandwidth, high priority traffic such as voice.
Internet providers would have to set that up, and to agree on a standard so everyone can latch onto everyone else’s voice-optimized infrastructure ?