The judge has ruled, and I accept his decision.  A little more than a year ago, I remarked on Ooma’s product and service, and declared that they would not last a year.  Dean Collins took me up a steak bet, and Alec agreed to be the judge.  I lost. Congratulations to Dean on his steak dinner, and to the employees of Ooma who have, so far, proven me wrong.

The basis of my bet was simple: the last thing that this industry needed was another residential box play, especially with the lose-a-little-but-make-it-up-in-volume strategy shown by Vonage. In addition, I felt like the value added feature strategy was ill conceived and basically ignored what years of value added services on cell phones have taught us: people don’t know they exist, and even if they do, they don’t tend to change their habits to use them.  And, speaking of cell phones, if there’s a clear thing about telephony it’s that we’re converging on the mobile device.  Try sticking one of those Ooma boxes in your pocket.

What did I get wrong? Well, this is still conjecture, and I’ve been wrong before (obviously), but here’s some thoughts:

1. I made the bet about “being in business”.  Apparently, it takes a while to spend $26 million dollars, and with their new $16 million round of funding, it looks like they’re going to be around for a while.
2. Maybe this is a case where, even though I was right relatively speaking, it’s a big, big world and there’s room for also-rans.
3. I should have predicted the arrival of top-rate talent like Rich Buchannan, who has succesfully placed Ooma into Best Buy (I saw it).

The bottom line? I lost.  The investor’s bottom line?  -42 million dollars, so far.  I still say I should buy Dean a hot dog, for this still isn’t a steak event.

To those at Ooma who might read this: I’m really not rooting against you.  Please prove me wrong; I’ll admit it when I am.   I’m just one who looks at the (insane) ROI that voice mashups provide, and see the 100 million dollar exits from Ribbit and asks himself “Why the heck would anybody throw 42 million dollars there?”.

Dean, if you would, please forward to me a picture of you in front of the steak. Make it a good one, and add a lobster tail.  Somebody deserves to make something from Ooma. Glad it’s you.

Congratulations to Irv and the Chicago based IfByPhone crowd.  Today, Information Week’s Serdar Yegulalp wrote up a review of the IfByPhone service – and loved it :

So far the award for Single Coolest Thing Done With Open Source at the Web 2.0 show has to go to Ifbyphone. These folks have used open source to make custom telephony applications as easy as designing a Web page. Easier, even.

Indeed, and what a nice thing it is to see the larger world of media paying attention to voice and phone mashups.  As a quasi-independent voice, IfByPhone is the best example yet of how you can add voice functionality quickly and easily to web pages. And for the web page designers, it’s all they might ever need.  Good job, guys!

Mashups featuring BroadWorks VoIP address disaster recovery, business issues

GAITHERSBURG, MD, September 18, 2008 - BroadSoft, Inc., the leading provider of VoIP application software, today announced the winners of its BroadSoft® Xtended Voice Mashup Contest. Winning entries include applications for disaster dispatching, accounts receivable and an Adobe® Air development mashup.

The Xtended Voice Mashup Contest challenged Web 2.0 and telecom developers to combine BroadWorks® VoIP features with other Web-based services including customer relationship management tools, business collaboration and social networking sites. Telephony applications such as SMS, voicemail, fax and mobile clients also qualified for submission.

“The ability to create mashups of Web services with voice features allows developers to quickly and easily create applications that positively impact public safety, business processes and other communication functions in ways that have not yet been possible,” said Michael Tessler, BroadSoft’s CEO and president. “Voice and Web 2.0 communication mashups are enabling innovation across all industries, allowing them to operate more efficiently.”

The winners are:

First place: Disaster Dispatcher
Created by Hyannis, Mass.-based Thomas Howe, the Disaster Dispatcher is a voice mashup that integrates Twitter, BroadWorks and RSS feeds to provide a one-window communications tool for emergency operators. By keeping all data in one window, emergency personnel can analyze the collection of information after the emergency to optimize response procedures.

“Disaster Dispatcher is a true mashup that integrates multiple services, a real use case, is Web-based and thus widely accessible,” said John Musser, founder and blogger of Programmable Web, a site that covers mashups, Web 2.0 APIs and the Web as a platform. “Twitter notification and RSS integration are creative applications.”

The first place winner receives $5,000 in cash and a trip to BroadSoft Connections 2008 with a free Solutions ShowCase booth. The package is valued at $15,500.

Second place: Project ARCTIC
Created by WorldXchange, a telecommunications carrier in New Zealand, Project ARCTIC improves the accounts receivable process within organizations. It leverages low cost methods to notify customers of past due accounts and ensures greater probability by providing reminders so bills are paid on time.

“Arctic is a well-conceived integration, surprisingly with a focus on the back-end side of the equation,” said Rich Karpinksi, reporter at Telephony Magazine and blog content provider for Telephony 2.0, covering news, technologies and business strategies of telecom providers. “The low-touch and low-cost automated feature makes it a nice customer service application.”

The second place winner receives $2,000 in cash and a trip to BroadSoft Connections with a complimentary pass. The package is valued at $5,500.

Third place: QuickSet
Developed by Las Vegas-based Joseph Farrar, QuickSet provides Windows and MAC operating system users with access to their BroadWorks services via a simple desktop interface. The interface is powered by the Adobe AIR platform which provides the ability to run Flash applications on the desktop.

The third place winner receives $1,000 in cash and a complimentary pass to BroadSoft Connections 2008. The package is valued at $3,500.

The winners were chosen from a finalist pool of six that were announced Sept. 10. The applications where judged based on their innovation, usability, benefit to end-users and their level of mash-up sophistication.

“We were pleased to be involved in the application judging process to provide feedback on what users want and can use,” said Dave Gilbert, contest judge and CEO of SimpleSignal, a BroadSoft-powered network. “Our customers are increasingly looking for products to make them more agile, and the winning products do so.”

The applications will be made available for download on the BroadSoft Xtended Marketplace, which was launched in March 2008. There are more than 1,000 registered users of BroadSoft Xtended.

About the BroadWorks Platform
BroadSoft’s IMS-compliant BroadWorks® platform provides a comprehensive range of VoIP applications, including Hosted Unified Communications, Mobile PBX, Business Trunking and residential broadband services fully integrated into a single VoIP application platform. BroadWorks is capable of providing these applications with the reliability, redundancy, scalability and regulatory capabilities required to deliver carrier-class service.

About BroadSoft
BroadSoft® provides VoIP application software that enables the delivery of hosted telephony and multimedia services. Its award-winning flagship BroadWorks technology empowers wireless, wireline and cable carriers to deliver next-generation voice and multimedia applications and advanced features that enable them to increase revenue, enhance competitive differentiation and elevate customer satisfaction. BroadSoft’s family of carrier-class software products delivers the scale, open architecture and reliability that the world’s leading telecommunications companies demand to serve mission-critical enterprise and residential broadband customers. BroadSoft provides VoIP applications to seven of the top 10 and 13 of the top 25 largest carriers worldwide, as measured by recent annual revenue, including Korea Telecom, KPN, SingTel, Sprint, Telefonica de Espana, Telstra, T-Systems, and Verizon. For additional information, go to www.BroadSoft.com.

Or, in other words, the real story behind Comcast’s issue with network neutrality.

I ducked out of my business meetings today to share a beer with an old friend; someone whom I’ve crossed technical paths with for nearly a decade. After talking about the structural issues that prevent Enterprise customers from getting what they really want from telephony, “Deep Bit” looks at me and says “It’s about the money. Just follow it, and you’ve got your answer. It’s just like the network neutrality issue and Comcast. You know about that, right?”

“Umm. Sure. I watch the Daily Show. I’m hip.”

“No, you understand why Comcast is whining about Bit torrent, don’t you? It has nothing to do with network congestion.” I stare. “The reason that Comcast is whining about Bit Torrent has nothing to do with network congestion. They don’t care about that crap. The real reason that Comcast is upset is because they have to pay carriage. Comcast leases out their backbone connection from Level 3 and others, and when people use Bit Torrent, Comcast has to pay for that. They could not care less about congestion, they’re just sick of paying carriage to others. The whole network congestion thing is a smoke screen.”

“Are you sure? I’m pretty sure they were complaining about network quality.”

“Here, listen. Who’s been throttling traffic? MSOs. Comcast, Rogers, Cogeco. Why? Not because they are afraid of network congestion. It’s because they don’t want to pay for those bits. When the SBC purchased ATT and Verizon purchased MCI, all of the contracts got recycled because of the change of control clauses. When that happened, the cable companies had to renegotiate, and now get their access from Level 3, Qwest, Global Crossing. They weren’t about to go back to the carriers, but the price they paid is that they don’t have Tier 1 access any longer. And that means that they are paying for those Bit Torrent bits, and they don’t want to. Why else would you see Level 3’s core network services revenue go up?”

“So, if the real reason for Comcast screwing with the bits is about money, and not quality of service, why isn’t anyone talking about this?” I cocked my head to appear cynical, but “Deep Bit” pressed on.

“I like my job. All the Internet architects like them too, and we’re not about to talk about it. We have families and 401Ks to worry about. It just pisses us off that Comcast wont’ come out and tell the full story. They don’t own their backbone, they pay for the bits, and it’s killing them. All this congestion stuff is bullshit. It’s about money, pure and simple. ”

Hmmm.

For the last two days, I’ve been attending the Serena Tag conference in Santa Clara. If you are unfamiliar with the Enterprise side of mashups, Serena is the maker of the Mashup Composer. The mashup composer is a tool designed to capture and deploy business processes at a level appropriate for the typical enterprise developer. For instance, imagine that you work in the human resources department in a typical large company. One common task is “employee onboarding”, where you assign them an office, a telephone, a mail box. Using the composer, you can design this business process using a flow charting tool, then publish it to be used through a web browser. The tool integrates seamlessly with all the usual suspects: databases, Active Directory, email, etc. If Ruby is my enabler for sewing together functionality, from phone calls to web service calls, Serena is my enabler for sewing together the business process.

In practice, most large companies have large and complicated existing solutions for the typical processes: compliance, inventory, returns, etc. However, expressed as the number of applications, most Enterprise applications are small and custom. You might only have a single Peoplesoft implementation, but you have a hundred smaller applications that you take care of as well. These small projects are often un-attended to by an oversubscribed IT department, and Serena’s tool allows the other work groups to quickly create their own applications. Serena is not looking to supplant CRM and ERP applications with mashups – they are looking to fill in the holes that exist in every company.

At the show, I’ve announced and demonstrated two mashups that I’ve created using Serena’s mashup composer. The first is an international wire transfer mashup, designed to allow the sender of a wire transfer to provide visibility into the progress of the payment using voice, text and email. The second is a re-write of the password reset application. I’m entering into a partnership with Serena, where I will offer my subject matter expertise around enhancing the business process using real time voice, and they will offer my mashups through the Mashup Exchange to their customer base. I’m very excited about this partnership, so look for some public announcements soon.

One more thing…. it’s very easy for us in the telephony world to lose sight of how large the Enterprise world can be. I’m certainly impressed at the names I’m seeing on the name tags at the show – nearly 1,000 people from around the world attending a show to learn about mashups and the future. For Serena alone, check out these numbers:

• 325 mashup assets
• 25 charter member and subject matter experts
• 2,000 Serena community members
• 15,000 Serena customers
• 96 of the Fortune 100 and 90 of the global 100, including the top ten companies in the top eight industries

For communications mashups, like the Disaster Dispatcher, one important question to answer is “How long do they take?” Now that I’ve done a few of these, and it seems like I’m doing them more and more and more…. I’m getting a sense for it. First, though, let me give the appropriate warnings and caveats: like all software, the schedules, budgets and risks rise exponentially with complexity. My mashups tend to be simple business processes with three or four different users, are integrated on existing platforms and web services, and use tool sets I’ve come to know fairly well. So, given that… how much effort are they?

In short, a fully deployed voice mashup takes around four weeks. This number assumes the size given above, and that the business process consulting piece is essentially complete before I start my work. Here’s the details:

• It takes about a week to communicate and understand the work of the
  business analysts, to put together a prototype of what you think they
  need, and for them to say “Yes, that’s what I wanted.”
• It takes about a week to tie up the implementation first version of a mashup. It’s pretty functional, with decent error checking, comments and help.
• It takes about a week to integrate the implementation of the mashup onto the existing Enterprise architecture, if it’s a project that requires it.
• It takes about a week to deploy it… training and testing, cutovers etc.

So, as a rule of thumb, it’s about a month’s worth of work to do it right. So, given that, where are the risk points? I’ve got some scars to share:

• The business analyst might know what he wants, but especially with enterprise software, there’s a cross functional team that wants their voice heard. You can very easily grow the analyst work for no other reason than organizational issues.
• Put aside at least a day for each Enterprise interface you need to integrate to, and make sure you think about unit testing for each one. So, if you need to integrate to ActiveDirectory – make that a day. If you need to integrate to some crazy stored procedures – that’s a day too.
• Like other web technologies like HTML, prepare yourself for the executive who says “I really didn’t like the tone of voice in that last question. Can you (re)record it again?”

But that’s about it. These rules of thumb hold pretty well, and maybe because people naturally resist audio applicaitons that are too difficult, or adding voice to an existing application is fairly straight-forward. As I look at the diaster dispacher work, it was written over a day and half, no more than two, but needs better documentation and error checking… so it’s about a week. No business analyst for this one (as I picked rather randomly), and no Enterprise platforms to deal with either.

If you have experiences – share them! It helps us all.

Now that I’ve had some sleep, let’s look inside the disaster to learn some stuff about voice mashups. First, let’s see how it matches up with what the hallmarks of a voice mashup are:

• Not a Voice Application : First, even though this is a submission for the Broadworks mashup contest, this isn’t a voice application. Voice mashups are not voice focused; they are application focused. This application is a framework to help a dispatcher communicate with field workers, report status to his superiors or outside parties, and to save what happened for later analysis. If you take voice out of the application, you still have an application. If you take voice out of conferencing… you’ve got nothing.
• Web as Platform : Secondly, this application is built on the “Web as Platform” model. The application runs on top of several other applications and platforms, and integrates them together using web services. In this case, we’re using Twitter as one platform, and Broadsoft as another. We are using an RSS interface to hook into another platform in real time – this could be a master feed from every dispatcher, or a feed from a major news service… whatever.
• Uses Web Technologies : Thirdly, this application leverages true-blue web technologies. Instead of using .NET (an Enterprise technology) and Java (an Enteprise and carrier technology), it uses Ruby on Rails as the framework, on top of web services interfaces to Twitter and Broadworks, RSS for importing data, and supplies an RSS feed to other data management services, HTML as the UI and Amazon’s EC2 as the hosting engine. It even uses super-hip, will never use anything unless bodily threatened, HAML as the markup language.
• It’s small. Real small : Mashups address the long tail of development, which means that the apparent audience for each one is quite small. This is self consistent: the power of mashups are that they are practical to construct from easy to learn technologies. Here’s a good visualization… think of a space shuttle. Got it? Now think of the opposite. Yup, that’s a mashup. Less than 600 lines of code here, and a good deal of it is UI. Poor Broadsoft has like, three lines of code or something.
• It’s focused. Real focused. : If you are anything but a disaster dispatcher, would you use this application? Unlikely. But if you are, I guarantee you that this tool, or something a lot like it, will be invaluable. But hey, no problem there…. since mashups are so easy to put together, you can make your Accounts Payable Autobugger, or Sarah Palin Flickr Imager, or whatever you’re into.

So, as you can probably infer from the application and the post, the real reason I did the Disaster Dispatcher wasn’t to dispatch our boys in blue to the flowing magma, but was instead was to help show what a voice mashup is, and how it’s put together. I suppose I really didn’t do it to compete, not that I wouldn’t love a trip to the Broadsoft Contest. So, I’ll touch on that one tomorrow: Disaster Dispatcher Technologies.

Well, at least I hope so.

Over the holiday weekend, I dusted off TextMate to send my entry in the Broadsoft XTended competition called “The Disaster Dispatcher”. Like most of my work, it’s business process oriented stuff, this time aimed at the emergency management market. I wanted to make this entry very light and fast, just like a mashup should be, and Gustav was all over the TV. (So was Sarah Palin, but apparently she wasn’t as inspiring. Must be getting old.)

So, I hereby offer my custom made voice-mashup, the Disaster Dispatcher, for your consideration. It is running live on Amazon’s EC2 at http://ec2-75-101-244-168.compute-1.amazonaws.com for your mashing enjoyment. This voice mashup is written in exactly 594 lines of Ruby, and integrates broadworks, Twitter and RSS to create an emergency services mashup. I wrote it over two days: Labor Day, when I wasn’t running errands with the wife, and tonight after work. I’m making this application open source to serve as an example of how to integrate Broadsoft and Ruby together to make an effective application. The complete source is now at github: http://github.com/howethomas/disaster_dispatcher/tree/master