GMail voice - nice, but no Skype Connect


So you can now make telephone calls from within Google Mail. Well, among other things this is a fine example of something we said back in 2008 in the Consumer Voice & Messaging 2.0 strategy report. Jamie Zawinski said that every program tends to expand until it can read e-mail - we said that the same was now true of telephony. Everything expands until it can place phone calls. As a result, although total minutes of use keep rising, the market is deconcentrating, with the total spread across an increasing diversity of players - games, Voice 2.0 companies, enterprise VoIP networks, mobile apps, perhaps even the odd telco.

But we actually don't think Google's move is enormously significant. Consider this: if you're a telco, and you provide plain SS7 circuit-switched voice, everyone agrees you've got a problem. Telephony is now a software application and it's very often free, which doesn't leave you much scope. If you're one of the traditional alternative voice providers - calling cards, carrier VoIP like Vonage, discount MVNO, etc - you also have problems, because you're trying to undercut a price that's going to zero. We recall Boris Nemsic, when he was CEO of Mobilkom, saying that their answer to "fixed-mobile convergence" was a new tariff that offered unlimited national and on-network minutes for €10. There wasn't any point being cute, when they could just cut prices and squash the margin players like bugs.

So you need to find some way to differentiate - to offer better voice and messaging.

Now, we were trying to find out if the Google product was actually Google Voice, which does provide a range of better voice features - visual, integrated voicemail, presence/availability, transcription etc. Hence the reference to Google voice rather than Google Voice in yesterday's news post. Here's Andy Abramson, who discovers that the new service brings you that fine telecoms industry product, a traditional IVR voicemail inbox that doesn't work with either your e-mail or with Google Voice's unicomms features. Also, the technology leaves something to be desired - it looks like they've built a JavaScript JINGLE client that logs into the Google Talk XMPP servers, but some users are reporting that it drops the call if the browser tab or window it's running in goes out of focus.

If some other company had launched a basic cheap calls service, we wouldn't have thought that particularly interesting. So we shouldn't just because it's Google. Sure, it means even more pressure on voice margins, but that's hardly news.

So what's Skype doing? Alec Saunders points out that the big news over there is that they've rebranded and relaunched Skype for SIP, now known as Skype Connect. The point here is that devices other than Skype clients can log in to the Skype network - indeed, anything that uses SIP can send a REGISTER to sip.skype.com and get online. That means, for example, that your company's Asterisk PBX could get its bandwidth from Skype, or your fleet of Avaya call-centre turret phones could log into Skype. Because both SIP and the proprietary Skype protocol support passing a lot of other messages before, during, after, and outside calls, your Voice 2.0 applications should work with it. There's also a new Skype Manager application to keep track of it all, and of course you can still use SkypeIn numbers, SkypeOut PSTN interconnection, and Skype click-to-call.

It's $6.95 a channel a month for all the minutes you can eat, and Skype has gone to the trouble of getting Avaya and Cisco (as well as quite a few others) to certify interoperability with their enterprise voice kit, so the IT department can rest easy. So it's a lot more than just cheap calls - although the calls are pretty cheap as well. What's genuinely impressive, too, is that they've clearly thought through the implications of doing SMB and enterprise voice. The announcement on the official Skype blog says, first up, that this is about connecting unicomms and IP PBX systems to Skype, and goes on to press the issues of certification, channel marketing, deployment, and technical support.

So Skype is lining up a major challenge in enterprise voice. Who makes an awful lot of money selling IP PBXes, unicomms gear, and desk phones? Cisco! Hence this rumour, which would see Cisco Systems acquiring Skype for $5bn. It's more convincing than the idea that Google voice with a small "v" is part of a huge strategy to destroy Facebook by encouraging people to stay logged into Gmail (the two things are not mutually exclusive...) And Cisco has been expanding its interests in collaboration services for some time - think WebEx conferencing.

It's harder to see what would be worth $5bn in such a deal. Certainly not Skype's margins on international voice. Perhaps the idea would be to boost sales of Cisco's higher-margin hardware. But Skype Connect means that Cisco hardware can already interoperate with Skype. There's obvious cross-marketing potential, but that doesn't need the ritual sacrifice of shareholders' funds on such a scale. It is true that Skype is on the hunt for major partners - Samsung's coming Android tablet may ship with a Skype client as part of the understanding with Verizon. And there's evidence that they're interested in a two-sided business model around click-to-call.

Again, though, the only benefit from burning $5bn would be to the Skype founders - which is a reason for them but not for Cisco.

The really big news for Skype, though, is the developer API coming this autumn. It's going to be absolutely critical to their future - as this piece on the Fring row makes clear. Can they really have been keeping Phil Wolff waiting for details of SkypeKit for the last 125 days?

Meanwhile, the question of Google's motivation has been raised. Most of the media/analyst chatter about this has been obsessed by Facebook. We consider this to be purely buzzword-driven analysis. Google has huge software development resources and a culture that prizes innovative projects; it's both cheap to do a small development project, and highly incentivised. As a result, they do a lot of projects.

Another possibility is that this is an effort to gather social-graph information from GMail users. However, it's very likely that the call patterns will be dominated by the names in the user's existing chat roster - the very fact it's integrated with the roster guarantees this. This is information Google already has, although there is probably a significant degree to which a typical user's IM and voice networks don't overlap. (Similarly, my GTalk and Skype rosters are similar but aren't identical.)

Also, remember when they served links in different shades of blue to randomly selected users, to see which shade got clicked more often? Google likes experiments. It maintains coherence by using the experiments to kill off the dodgy projects. We believe this to be another experiment - perhaps there is some interesting data to be had? Or perhaps Google is just trying to learn a little more about this odd "voice" thing?

Either way, until and unless they develop the sort of understanding of voice & messaging that Skype Connect demonstrates, it's mostly another cheap calls offering with some extra features. Google could deepen the integration of GMail with Google Voice, and Alec Saunders' post referred to above offers some valuable pointers - it remains true that cheap calls are no longer enough, and there's a real market for better voice and messaging. Ask Skype.

We'll be discussing how to manage the co-opetition with Google, and the challenges of Voice 2.0, at our next Telco 2.0 Executive Brainstorms, in LA on the 26th-27th of October and London on the 9th-10th of November.