Thanks to everybody who dialed in for today’s Voice Mashup conference call, especially to Jim Courtney, Andy Abramson and of course, our host, Alec Saunders.   We talked about what a mashup is, how we make money at them, why are they written and who (in general) really cares.  If you didn’t happen to be one of the people in the audience, no big deal… you can listen to it over on Alec’s blog.     Dan York’s experience wasn’t entirely positive, as his hand was raised and the teacher didn’t call on him, but I thought (on the whole) it worked out beautifully.   On a personal note, I couldn’t help but feel the benefits of the whole “social networking” thing from the process of doing this call.  There are many reasons I don’t often write on that topic, not least of which is that I’m pretty darn focused on mashups, but I found so many good points about this experience that I just had to share:

1. On the message wall, Luca chimes in with his monetization insights for mashups, and  Dean reminds me of our steak bet that Ooma is dead meet in twelve months.
2. I can share it with you guys by linking to the podcast itself.
3. I saw people who I  added to my Facebook friends list by looking on the attendees to the call.

I’m often asked to give an examples of a business case for a deep integration of the business process with telephones, so here’s a good one for you: Morisky Surveys.

Jott is a true diamond in the rough.  

For those of you with significant others, I don’t recommend that you share this information. Not that I have anything to worry about.

I was lucky enough to catch the end of Teltech’s presentation at the Cluecon 2007 show in Chicago, and I heard about their new service called LiarCard. Essentially, Liarcard listens in on phone conversations, and detects if one of the parties is lying. Honestly. (Oh God, could this get bad, quick.) You arrange a call through Liarcard, and it records the conversation and then uses voice analysis to determine the probability of dishonesty during the call. You can even go back afterwards and tell which parts of the conversation were more dishonest than others.

Is it accurate? Well, according to them it is :

Essentially, if the quality of the voice is reasonably good and the operation and preparation is proper, the emotional analysis component will be almost 100% accurate. In this case, the technology will properly present how the tested subject is feeling in terms of emotional charge, cognitive conflicts and general stress (“Fight or Flight” syndrome). If the intention to deceive is genuine and this poses jeopardy on the tested subject, (assuming the tested party falls within the standard range of “sanity” or “normality”), then the Inaccuracy and Lie determination will also be accurate more than 90% of the time. (In the latest field research study conducted on 500 passengers in an airport, LVA -the security version of the technology- was able to render an overall accurate analysis in all 500 cases.)

Ok – I’m not sure if this is the most amazing service I’ve seen, or the most disturbing, but I’m sure that Liarcard could read my emotional response. Hmmmm… I see they also have a site called LoveDetect… I wonder what THAT’s about.

Anyways, I normally give out report cards for places like Liarcard, but I think I’ll choose my words carefully this time. I’m sure you understand.

According to BCC Research, spending on clinical trials in the United States was almost $24 billion dollars in 2005. According to Microsoft:

Pharmaceutical companies can spend 12 to 15 years and up to $900 million to bring a drug to market. About 45 percent of this cost is accrued during the clinical trial phase. Additionally, studies indicate that 75 percent of all trials conducted in the United States are behind schedule by one to six months. Because improving time-to-market for new drugs is critical for pharmaceutical companies, managing the clinical trial process is one of the most significant areas of opportunity for improvement.

Critical to the success of any trial is the consistent, streamlined and reliable collection of patient data. Exacerbating the problem are logistics, for most trials involve hundreds and thousands of medical personnel and patients. Pharmaceutical companies must leverage technology to help teams communicate and to collect patient data not only for cost reasons, but quality as well.

I believe that this problem begs for a solution based on programmable web technologies. Using the phone as a input device, patients involved in the trial can give consistent feedback that is instantly available to researchers. Using the web as a platform allows for simple and reliable integration with existing equipment in the phamaceutical vendor’s systems, especially when issues such as geography or inter-company communication are involved. Using technologies such as Voice XML and Ruby on Rails, reliable and scalable systems may be custom designed to collect data from patients in very rich ways, decreasing the time to analyze results from trials, speeding time to market, while lowering costs.

Jaduka is a web services company aimed at allowing Mashup developers to trigger phone calls from web applications. Essentially, if you want to deploy a button on a web site, so that customers can click and call you or your business, Jaduka has a great offering. I put mine together in about five minutes; you can too. I picked the button I liked the best from their copious gallery, and off it goes. It’s working, and I’ve pointed it at my Grand Central account, if you want to try it. I picked the button below, as I bet Jon Arnold will call me and I don’t want to have him dive for his headset again :

Jaduka also provides an API, which has two basic parts. The first manages calls, either calls to the account holder’s phone number, or calls between two arbitrary numbers. The second part of the API manages voice mails, so that you can manage them as wave files. As an example of the call management solution, you easily bridge calls between two numbers whenever it makes sense. Let’s say you have a scheduled call with a partner. You could make an application that calls you when the appointment starts, then calls your partner. You can also make that functionality point to different places, so that you can implement a find-me, follow-me system, or a skills based routing engine. As an example of the voice mail management solution, you could take your Jaduka account and aggregate reports from remote salesmen by having them call into your number, then taking those voice mails as WAV files and attaching them to your CRM system.

Finally, let me say that whoever is doing product management at Jaduka is doing a great job, because the Jaduka API is small, easy to learn, and provides real value. I didn’t have to spend more than fifteen minutes to read the documentation before I felt like I grokked it. Definitely part of my toolbox going forward.

Just because you could, doesn’t mean you should. If it is much easier to blend in speech recognition, IVRs and SMS messages into your software, should you? My theory is that there are three positives from blending real time communications with the business process : it makes businesses faster, it makes businesses more efficient, and it makes customers happier. Let’s look at an example that shows each.

Many delays in business are human delays. Finding the right person to answer a question, sign a form or move a phone line can take as long as doing the task itself. Integrating real time communications into the business process makes the job of finding the right person a matter of software, not people-ware. My example for making businesses faster is the review cycle for reports, designs or other plans. If the review cycle were handled by a piece of software, it could coordinate all the comments, aggregate them and post them for the team to see. It would be faster if, as the deadline neared, the software would actually call your cell phone and ask “Your attention is required on a review due tomorrow. If you have any comments, just speak them now, and I’ll add them as text to the review document. If you don’t have the time to review this, say ‘I don’t have time’. If you need 30 minutes to review the document, say ‘Call me back in 30 minutes’ “. Reviews would happen quickly and predictably.

Keeping with the example, I think it’s easy to see how this makes business more efficient. How many times have you been to a review meeting, which for most people for most of the meeting, a complete waste of time? Yeah, me too. How about the time to coordinate the meeting? The time to aggregate comments and post them to a site, if you ever did? Answer this honestly: can you go back and get the comments for any work you had reviewed two years ago? Integrating real time communications into the business process makes the business more efficient by coordinating feedback, by making it easy to forward comments back and by (frankly) nagging appropriately.

Happier customers? Yes, I’d say this qualifies as well. As a manager, and the customer of this process, I would be happier to see a predictable schedule, with a well formatted and organized output. Our studies show that one way to make customers happier is to increase visibility into the process itself. If you can see what’s going on, you can make judgments about progress and success, and that will help you relieve your anxiety. Same deal here, as you can see how the review cycle is moving, who has responded, and who has not.

I think a fair question to ask would be at this point, “Well sure, Tom, but I could have done this with Lotus Notes and a workflow.” Yes, but I think the addition of voice and real time communications has made it even better, and valuably so. By giving the option of offering a review over the phone means that I don’t have to wait for the boss to get back to get feedback, even if it’s “I don’t have any.” By giving the option to nag your teammates for their feedback means that I don’t have to hope they read their email. For those outside your company, I don’t have to fear the SPAM filter. Never mind the nuance of emotion that can be communicated easily with voice, and gets lost with text. Blending in voice to does this for you.

This is a horizontal application, but nearly everywhere I look I see the same effect. Disease management has similar issues, credit calls, transportation, inventory and purchasing as well. You may claim that these are small issues, but I would disagree. But even if they were, costs of deployment for these technologies have fallen so far that to deploy these solutions won’t break any company’s budget.

“OK, so let me understand this… you guys do IVRs, right? I don’t really understand what’s new here. We’ve had IVRs forever.”

To live in the real world means to live with constraints. I am unable to jump over a tree. I will live less than 100 years. I am unable to add numbers as fast as my computer. I cannot fit three cars in my bedroom. As we move through our lives, we make decisions based on these constraints so often, we no longer understand that we do it. Good thing too, because if we did have to run through the list of what we couldn’t do all the time, we would be paralyzed. (I have to be in the Cape Cod office in 90 minutes, so there is no time to swing by Paris for breakfast.) Our minds are honed to consider the possible by a clever method: we tend to construct our understanding primarily based on our current knowledge, giving rise to an endless series of deltas. I remember an English class where we were working on a dictionary entry formula : you define a noun based on two parts, the first being what the noun is like, the second is how it’s different. A tiger is a 1) CAT that 2) has stripes, a big body and large fangs. You define a tiger using what you know (a cat) and then describing why that first thing is insufficient (which can eat you for lunch.)

Of course, engineers are no different, except that when we design, constraints are quite at the top of our attention. I would even posit that we understand and describe our designs through constraints. Like a sculptor, removing the stone to reveal the statue, we draw boxes on whiteboards to explicitly limit the functionality of what we design. We understand that, unless we limit our tasks aggressively, we will be unable to see our design come to life. It why we cringe when the marketing man comes in the room, as his motivations are quite different than ours. They make us draw more and bigger boxes, removing our constraints and adding to our worries.

What might not be apparent from the outside is that an amazing has happened recently in the world of technology. The combination of web services, lightweight programming models and Internet architectures based on open source tools radically reduces the constraints on hosted application development. The advent of Web Services means that you can publish your functionality to the Internet in a controlled, standard way. Since the cost of publishing is so low, the amount of customer traction you need to break even is low, making the amount of web services that are available to you as a designer to be much larger. An example would be a web service that verifies that a person’s name matches an address, and will be valuable to any company looking to avoid fraud. Lightweight programming models radically reduce the cost to create web applications by reducing the skills required to write them, and the number of people required to author and maintain them. Internet architectures are naturally resilient and scalable, removing many of the stability constraints that dog other forms of software development. As a concrete example, I would have you consider that both Yahoo! and Google run their infrastructures on off the shelf hardware and open source software, versus Verizon who runs their infrastructure on gold plated, multi-million dollar telecommunication switches. The difference is in architecture, which removes the constraint of having each element be bullet-proof and stable. In the Internet architecture, servers can go down because there are multiple paths to your web page. In telecom architecture, the phone on your wall is connected to exactly one central office.

The combination of these forces makes the development and delivery of hosted services much, much less expensive, for services of all kinds. My career interest is in telecommunications, thus, I am looking at how the constraints of typical voice services are lowered. IVR is a good example, of course there are many others. Using an online service from a place like Voxeo, I can draw a box on my white board called IVR that I can implement quickly and inexpensively. How might I use it? In any way I wish, blended into any sort of application I wish. I am no longer constrained by having to purchase equipment, or by having to learn an esoteric language to run it, or to be locked into a particular vendor, or to have troubles interfacing it to my software. It is no exaggeration to say that I am now able to make function calls that retrieve data from an IVR script as easily as I can that retrieve information from a database. In fact, given that it takes me a couple of hours to setup a database, the IVR might be easier. I can tell you the same thing about a whole host of telephony services, like SMS. I could also make the same argument about any other web service used from my IVR, like I could call a phone number and make my google maps on my screen do things if I wanted. I could – any up-to-date software engineer could.

So, given that these constraints have fallen through the floor, what does my company do? We are figuring out exactly which hosted solutions, especially those that use real time communications, work inside businesses to make the business faster, more efficient and makes their customers happier. Then, we work with businesses to implement these new services into their business process. IVRs are a part. So is SMS. So is Google Maps. So is…. you get the point. What is different is not the functionality, but the constraints. The implications are endless.

One of the best pieces of advice I’ve ever heard for software engineers is “Use the tools”. Like words for an author, tools help define how you think, and how you develop and ultimately, what you can develop. One of my best tools, and one that I find myself relying on time, and time again, is Voxeo. Voxeo is hosted provider of IVR solutions in Orlando, run by a really decent and talented group of people led by CEO Jonathan Taylor and CTO (a geek’s geek) RJ Auburn. Using Voxeo, it’s really easy to throw together IVR scripts and deploy them for demos, betas and production. And, I can integrate the results from, or use it drive data to, database driven web sites. Voxeo handles inbound and outbound calling, and integrates easily with standard and VoIP phones.

I’m actually doing a mashup right now using Voxeo that (I hope) is ready for Cluecon in a few weeks. I’ll post the details when I’m done, but here’s the idea. The Cluecon audience is pretty dense with geeks, so for fun, I was wondering if I could find the geekiest one. I’m writing a Voxeo script that will call each attendee (or they can call into it, if they want), and give them a survey. For instance, I might ask them if she-bang was a Unix script thingy or something that Gomer Pyle says. I’ll take all the answers, profile them, and find the person who’s answers were closest to the norm. Maybe we’ll even find the person who’s answers are furthest from the norm, and give him the Prom King award or something. Anyways, using Voxeo, this demo is pretty easy to put together, including the database site, and won’t take me more than a few days to put together, host and test.

RJ – I’m afraid you’re not eligible for this contest. I’ll send you an O’Reilly book as a consolation prize.

In sales, what kills deals? Time. Microsoft just enabled more sales to see the light of day.

I read today on Andy’s blog about the recent partnership announcement between Microsoft and Verizon. Sponsored search on Microsoft’s Live Search will now include a link that enables Click To Dial between the business and the browser. Since this is only available for sponsored search, I wasn’t able to check it out, firsthand. From the article, it seemed as though the partnership provided free calls from the business to the desktop, as opposed to free calls from the business to your handset. So, how does this stack up?

From a business perspective, I think it makes sense for all parties. I play this game with mashups all the time… take a new light-weight application, and ask yourself these questions:

1. Does the service make the business faster?
2. Does the service make the business more efficient?
3. Does the service make customers happier?

In this case, I think they’ve got three for three. The customer gives his permission, and gets
instant satisfaction. Not bad. More so, they’ve just radically lowered the time it takes to make the sale. The business makes an outreach to a self-qualified customer that’s waiting for the call, and it can’t get too much more efficient than that. And the customer is in charge of the interaction, its schedule and the mode by which it happens. In charge customers are happy customers.

From technology view, though, I have some worries. I think the computer makes a crappy telephone. (Sorry, if I offend all of you Skype lovers out there, but I gotta call it as I see it.) Where’s my headset? Click, click, ring ring… it’s a very aggressive and unfamiliar interface, for many, it not most, people. I’m sure my mother would be quite confused if the computer started talking to her. What does she speak into? All I can imagine is Scotty in that Star Trek movie where he picks up the mouse like a microphone and starts talking into it. No guarantee that the person on the other side of the phone knows that you’re calling from a browser… he’ll have fun trying to explain that one. Yes, there’s a certain percentage of users for whom the computer is a wonderful communications tool – and I suppose I’m one of them – but as a broad based endpoint? I’m unconvinced, and that’s after my fourteen years of VoIP softclient development. If Microsoft only depends on a browser based approach, I’m thinking that the upside is limited.

What makes a good telephone interface? A telephone. With mashups, that’s an easy fix. Click on the link, ask the user for their phone number, and ring it. Simple and easy. You could even have an option to use the browser for those 5% of our population that would actually prefer it.

What I want to know is how and when Microsoft will start mining all this call data. I suppose there’s an off chance they won’t, but only because they are still a few years behind Google. They will, at some point, to great advantage to themselves and their advertisers. Perhaps that’s why they (apparently) are going for the browser play, so they don’t have to deal with all of the call recording legal issues that vary per state.