For 40 minutes last night I had Ooma CEO and Founder Andrew Frame on the phone. I didn't get to talk with Ashton Kutcher, like Mike did, but that was okay. Ashton wasn't going to add anything to this conversation. My goal was to get some of the many technology questions about Ooma answered. In advance I forwarded a list of questions. Andrew answered them, and more.
Let's begin with the technology itself. The Ooma Hub and Ooma Scout are slightly different from what most of us in the VoIP world would refer to as an ATA. The Hub is a QoS enabled ethernet router, plus a media gateway, plus a HomePNA router, and an ATA. The Scout is a HomePNA ATA. For those not familiar with HomePNA, this is a 10 year old LAN technology that implements a high speed LAN (up to 240 Mb/s) on home phonelines. The important characteristic of HomePNA is that it modulates the network traffic on the phoneline outside of the audible spectrum. The WAN technology we all know and love called DSL uses the same technique. What this means is that the phonelines in your house can simultaneously be used as phonelines and as a high speed QoS enabled LAN.
An Ooma Hub connects to your broadband, your POTS line, and and an FXS connection for you to connect a handset. The POTS connection is used simultaneously for inbound / outbound POTS calls, and to provide the HomePNA network link for the Scout. The Ooma Scout is simply an ATA with a phoneline port rather than an ethernet port. Calls made between an Ooma Scout and the Ooma Hub are VoIP calls running on the HomePNA LAN. That's how they implement their "instant second line" feature.
When you pick up the handset attached to the Ooma Scout, you hear dial-tone, even if the other phone is in use. What a novel idea! When you pick up the phone, most often you're intending to make a phone call. So play dial-tone, rather than interrupting the other call in progress. You have the option to dial a call, or to join the other call. Conversely, when a call comes in and one of the other handsets is in use, all the other "line" phones ring.
Theoretically you could have an unlimited number of "second lines", constrained only by the bandwidth to your home. Andrew allowed that this was possible, and then explained their decision to limit it to two as usability and market need considerations. From a design philosophy point of view, consumers already know how to use flash hook to switch between calls when another call is waiting. From a market point of view, many homes have a second line… few have a third. It seems to me that this also opens up the potential for an Ooma "Small Business Edition" in the future as well.
The Ooma team has done a number of things to address quality. The Hub itself does all the things you would expect that it might. It's a QoS router. It uses traffic shaping techniques to ensure that when someone in the house is making heavy use of the WAN connection there is still enough bandwidth for voice. Because Ooma is a closed system, it can use it's choice of broadband codec (they're using GIPS) for voice quality. And perhaps most importantly, it was designed by a team of engineers who have done this before. According to Frame, it wasn't a case of "how do we build this?", but rather "how much time do we have?". Unfortunately, the ultimate quality constraint is the broadband provider. If necessary, however, this is where they can fall back to the POTS line. Users will know that they're making a POTS call because the Ooma dial tone has been replaced by a regular dial tone.
The feature of Ooma which generated the most commentary in yesterday's announcement was distributed termination — the peer-to-peer aspect of their product. If a third party is calling into my Ooma to reach someone in my local area, what caller ID is displayed? Frame informed me that the correct caller ID is displayed (that of the third party calling), but demurred on how it was done. "That information is proprietary", he said. Similarly, what prevents an Ooma owner from simply putting a phone on the home line, outside the Ooma, and listening to outbound calls? Again, Frame asserts that they have implemented solutions to the problem, but that the implementation is proprietary.
The success of their distributed termination strategy is dependent on other Ooma users opening up their POTS lines to the network. Ooma will provide terminations also, for areas the the Ooma network doesn't cover yet. However, these will be at a cost. What that means is that to get the benefit of their network, there will need to be a growing base of users using Ooma. Early adopters may find Ooma a better experience than their current phone, but calling costs will not fall until the network achieves some critical mass. They have about 250 users on the beta now, and expect to grow that quickly to 1500 or more with their white rabbit program. Frame believes that with 2,000 Ooma users they will cover 90% of the United States, and that at that point the network, capacity, and coverage problem will be solved. Intuitively, this makes sense. Only Ooma users can access other Ooma users devices. Once the network graph is largely complete, then every time a new user is added in a particular area, their added "capacity" adds to the network. One could argue that 2,000 devices are not enough, given that there are over 3,000 rate zones in the US, but the logic of how the capacity grows makes sense.
One of the other comments raised yesterday was that the Ooma distributed termination model changes and perhaps breaks the telephone experience that consumers are used to. I would have to argue that this is a fallacy. Disconnected from the POTS line, Ooma is a sophisticated ATA that connects to a VoIP network. Connected to the POTS line it adds 911 to the VoIP experience, which consumers depend upon. It also allows that line to be shared, and Ooma asserts that the sharing is completely transparent to the owner of the Ooma device. If anything Ooma is closer to POTS than most other VoIP solutions out there, with the potential exception of PhoneGnome.
One of the biggest questions in my mind was Out of Box Experience. How easy will it be to set up? Using some patent pending technology, they are able to detect and setup the Ooma without any assistance from the ISP. According to Frame, the Ooma router is completely self configuring.
Like PhoneGnome, the Ooma Hub is also a platform for new services. So, for instance, they deliver a sophisticated voice mail service with call screening. Initially, their focus is on replicating the experience that people already have on the phone, and then adding new services. The strategy makes sense — consumers won't accept an expensive product which does less — but I can't help thinking that Ooma needs a few innovative features to really make it a must have for early adopters. Currently, Ooma has no plans, however, to open the platform to
third parties. In this regard, PhoneGnome has them beat. In fact, PhoneGnome CEO David Beckemeyer issued a "build your own Ooma challenge" yesterday to any developers who want to try to beat Ooma to market. Nice!
At this point, after having politely put up with my technology questions for 40 minutes, Andrew Frame had to dash.
I'll be getting one of these devices in the future, and will write more about my experiences setting it up and using it. Unlike most of my friends in the VoIP blogging world (who've given this a resounding thumbs down) I remain open minded to the idea that the Ooma team has addressed the negatives. I am impressed that they're confident enough to put this in the hands of the ordinarly merciless Walt Mossberg. They must be doing a few things right. My biggest skepticism about Ooma today is a marketing issue, rather than a technology issue. Like Om Malik, I think they're deluding themselves that they will achieve significant adoption with a $400 price point. Expect this to drop to under $100 pretty rapidly.