Interview with Mark Cromack, Cogi

Story by Benjamin F. Kuo


Santa Barbara-based Cogi ( recently launched a phone-based service which allows users to transcribe conference calls and meetings. The firm's founders are all experienced entrepreneurs, and come from Callwave, CrystalVoice, and Ideocore. We sat down with president and co-founder, Mark Cromack, to hear more about the company.

For readers who haven't heard of Cogi, can you explain what your service does?

Mark Cromack: First and foremost, the concept behind Cogi is brought out in its name, which is short for cogent idea. The idea is that in a conversation, there are certain highlights and points that could capture the important parts of a conversation. That's what the Cogi service offering is all about--giving you the ability to capture, review, and share the content of your important conversations. At a high level, what that means is you can capture the content, record it, and also capture what you believe is the important parts of the conversation. So we're going along, and you say something that really sparks my interest, and you can push a button either on the phone or computer to let Cogi know this is important to me or important to you. I can review it myself, both in audio form, or text form, and share it with others. The share feature is the real differentiator because, again, what it allows me to do, for example, Sarah Savage, our marketing manager, wants to be able to understand what the questions were for my interview this morning, and she'll be interested not only in what questions you asked, but what my answers were.

You've got an interesting team, can you talk about the background of the founders?

Mark Cromack: Certainly, we've been in the telecom industry for the last twenty or thirty years. We started a variety of different companies. Our founding CEO, Bob Dolan, founded ComDesign and started Callwave, the Internet based service for delivering Internet call waiting. If you remember, the challenge was, if you were on dial-up, and the line was busy, there was no chance you could get through to that particular person. CallWave streamed the audio down from the telephony servers to your desktop, so you could hear if you had a really urgent message from so-and-so. The previously company I started was CrystalVoice communications. CrystalVoice solved the problem that plagues Internet voice applications. So if you were using Skype, Google Talk and had a call over the Internet, you'll notice in some instances especially when the Internet is a little dicey, the audio quality degrades pretty substantially. As the name implies, CrystalVoice solved that particular problem, and was able to provide a consistent audio experience in the presence of significant Internet challenges. That company was sold in 2007.

How much technology automation is involved here, or is there just lots of labor involved?

Mark Cromack: It's a rather complex process, because, first of all, you've got to get all of the telephony components together. You've got to have Cogi in the middle of your call. There are smart phone applications coming out, there are in-bound virtual telephone number service capabilities today, and there is a web gadget software solution--which I'm using now--which allows me to visually see what's going on, on a particular phone call. In this particular instance, I'm calling you on a regular phone, and talking on a regular phone, so it's no VoIP. But, it's still the marriage of telephony and the Internet, to create a seamless experience and ease of use for the subscriber. It's a powerful, visual tool for a subscriber to understand what's going on.

In terms of the underlying transcription, it's really a combination of three different things. There's a speech recognition component, and a quality assurance component, because the technology available today is not sufficient for the challenges of a day-to-day, business call. Imagine you're in Southern California in your car, you're driving down the 405, and t here's noise in your car, and you're on your cell phone--it's not the ideal situation for speech recognition. If you've ever used any of the commercially available speech recognition engines built into your car or into your cell phone, they're less than perfect. If you can imagine extending that concept into a full blown conversation, and where you're literally trying to get a word-for-word transcription of what's said, that technology is not going to be sufficient. So, there's a human element involved so we're able to correct those in an efficient cost-effective way, and so we're able to deliver the service in a cost-effective way. And, we have our own custom signal processing technology that increases the efficiency of all those existing systems that I've just described.

What are the barriers to entry here--it seems like there have been a number of web startups who are doing some aspects of this--if not conference calls, at least transcribing voicemails?

Mark Cromack: It's pretty substantial. Before we talk about barriers, let's talk about market applicability. When you look at companies doing transcription for voice messages, at 3000 feet, they seem really similar. But, if you look at a call that you missed, there's not a lot of content and it's very short in duration. But, if you're going to be recording an indefinite amount of media, you have to get in the middle of the conversation, because you're not integrated with existing voicemail. You need to be able to take the audio, and process it by whatever means is appropriate, and also provide an email or web solution to provide that transcribed content to the user. In the case of Cogi, we're in the middle of the call, and are dynamically transcribing an indefinite amount of content.

How's the company funded and backed?

Mark Cromack: We're privately funded, both through our management team and private investors, mostly in the Santa Barbara area.

In your business model, it looks like you're charging for the service. It's been popular in recent years to have "freemium" services--can you talk about some of the decisions you made to go straight to charging for the service?

Mark Cromack: I think the challenge is down the road, and what happens downstream. If I offer this for free, generate lots of buzz, and gets hundreds if not thousands of users, what have I validated? Do I know how many of those people value the service? Often times, research shows that when people pay nothing, they think that's the value associated with the service. Even before we began the process of starting work on our software, we did extensive market validation. We call it customer discovery--we went out to talk about the problem of conversational memory retention. Wouldn't it be great to have a button where you can have a transcript of your entire call, or for most business applications, the most important parts? When we presented the problem statement, this particular problem resonated so strongly with most people. We then would follow up, asking--what would you be willing to pay for this service? More than 90 percent of the time, the answer was substantially more than what we're charging with our introductory offer, $29.95.

How far are you in terms of launching and offering this?

Mark Cromack: We launched the service with a modest amount of PR in December of 2008. The company was founded in March of 2007, so we'd been developing the software and putting together the infrastructure for delivering the service since that time. We had an alpha as early as a year ago, and because the concept was rather novel and new, we needed to experiment so we could understand exactly what our customers would eventually be seeing. We began in earnest in the middle of January with a customer acquisition campaign, which has been encouraging.


More Headlines