Why Transcribing Your Content Improve Your Business, with David Feinleib (Speechpad)

Watch this great chat with David Feinleib CEO and Co-Founder, Speechpad!

We’ve been talking about transcriptions, how a good transcription can affect your SEO, how it works behind the scenes and much more (if you’re interested also in “Big Data” and “Quantified Self” keep on watching..).

About David:

Davis is an investor, advisor, and serial entrepreneur.

From 2009 – 2011 he was a General Partner at Mohr Davidow, a venture capital firm with $2B under management. David led the first institutional rounds in RootMusic, the #1 entertainment app on Facebook; doxo, the premier digital filing cabinet; and VirtuOz, the leading virtual customer service company. RootMusic recently closed a $16M follow-on financing; doxo, a $10M round, and VirtuOz a $7M round. He was a Principal at Mohr Davidow from 2006 – 2008.

Prior to joining Mohr Davidow, David started four companies, one of which was acquired by Hewlett Packard, another by Keynote Systems. Before that he was a Technical Evangelist at Microsoft, promoting the then unknown Windows operating system.

In 2011, his love for entrepreneurship took him back to working on his own ventures, while continuing to advise and invest in others. David hold an MBA from the Stanford Graduate School of Business and a BA Summa Cum Laude from Cornell University, where he was a Kodak Scholar. David is an avid triathlete and violinist.

Raw Transcription:

Marco: Hello everyone, Marco Montemagno here, the Tech Alchemist and today with me is David Feinleib, the Co-Founder and CEO of Speechpad. How are you doing?

David: Hey Marco. How are you? Great to be here.

Marco: As I told you, I have the most horrible surname to pronounce. Montemagno is something impossible to pronounce. So, I’ve been asking David before it began how do I pronounce finally-because it sounds like a European surname or you’re from U.S. What’s your . . .

David: Yes, It is from Europe. My great, great, great grandparents moved here, and we’ve been here ever since. That is where it is from.

Marco: All right. So, Dave, today with you I’d like to talk about several topics. I would love to start from this transcription world where Speechpad is doing its business now. For the community following the Tech Alchemist, they know that every time I upload an episode, there is a full transcription. That transcription I’m using it, both in the post and also inside You Tube to get some titling. It’s very, very useful, and there are several reasons I want to talk to you today, Dave, about why transcription is useful.

I’m doing that transcription with SpeechInk, so I’m in conflict. Interest conflict. I love to talk about product and services that I’m using. I’ve been doing interviews with user testing. All the services that worked, I think it’s very good to share that knowledge because other businesses can have the same need. Let’s talk about transcription market first. What are we talking about and why did you start with Speechpad? How is the transcription industry in this moment the state of the art?

David: We started Speechpad because we were experiencing bad transcriptions ourselves, and we wanted to fix that problem. Actually, we were getting these voicemails transcribed, and you would think even that, you could get really high quality machine- based transcriptions for that. But it turns out that it is still quite hard to get a super-high quality transcription, especially when there’s background noise, accents, multiple people speaking, things like that. That’s why we started it.

We looked at the market and it turns out the market is a $20, $30 billion dollar-plus market opportunity. It’s an industry that’s been around for a long time. It’s been quite hard to get transcriptions done easily and quickly over the Internet. That was kind of our premise. A really big market, it was a problem that was very personal to us, and we thought that there was an opportunity to disrupt it with a great Internet-based offering.

We started prototyping in 2009. We spent a quite a while prototyping the software, and then we really started scaling it up about two years ago is when we really put some muscle into it.

Marco: Right. I remember, I don’t know, five years ago, seven years ago, actually I remember when YouTube launched. The funny thing is that I remember the first videos that I’ve been doing, and I needed a transcription, what I was doing was finding a freelance, and then he was transcribing everything. Then, I found someone editing the video to put all the subtitling, and it was, I don’t know, one month to do the job. It was long and painful. When these kind of services, like Speechpad, started to come out, I think you guys solved a huge problem. The feeling of getting a transcriptionist is not so horrible anymore, but is a smooth process where you just upload a video and so on.

Can you explain to us, for people who don’t know Speechpad, how does it work from the beginning to the end?

David: Yeah, absolutely. We use a mix of humans and some computer- based technology to deliver the high-quality transcriptions that we do. Essentially, you can record an audio or video on your iPhone, your Android device, you can record it on your PC. You can record on the phone. However you are recording it, some people also do professional video production, and you know, record things that way.

So, you’ve got your audio or video, you come to our website at speechpad.com. We make it really easy to upload audios and videos in just a wide variety of formats. You upload your files. We take those files and send them out to thousands of transcribers. We literally have an internal system where we have thousands of people working on transcriptions. They do the work, and then it comes back to us. We proofread it, potentially do a little bit of editing, and then that transcript appears in your account and on the website you get an email notifying you that it’s there.

We also, of course, support FTP upload and a web services upload for customers that are really high volume.

Marco: Described like that, it sounds very easy, you know, from my point of view because typically I just upload a video on YouTube, go on Speechpad, copy and paste the URL, then I say, “Yes, Order,” and then I download my transcription one day or one week, it depends. You can choose the timing, right?

David: Exactly. Customers can choose 24 hours, 48 hours. They can choose a one-week turnaround. Obviously, different prices for those different turnaround times.

But I think one of the big things we focused on was, “How can we make it really, really easy. Where it truly is what you are saying.” If you upload your audio or video, as far as the customer is concerned that is all that you have to do. Then, you get a really great transcript back in the timeframe that you requested.

Behind the scenes, we are doing a lot of work obviously. We are spell- checking. We have rules. We are looking at the files. We have audio and video conversion, so that all of our transcribers can play the audio or video, different speeds, things like that depending on what you uploaded.

Again, that is all behind the scenes. We try and make it really just easy for customers. Audio and video in, great text out.

You can upload audio and video in many different formats. There could be MP3s, Wave files, WMA. It can be different video [codecs]. One of the problems of where we saw for the customers was just the variety of audio and video formats that they were working with that are in this space. You’ve got to pick a very specific format to work with.

Whereas with our site, you know, we really focused on making audio and video acquisition. That’s the process of getting the audio and video into the system really easy. So the customer can just take whatever they have and get that into the system, and then we take it from there.

Marco: How could you recruit so many people willing to transcribe? Because if you asked me to transcribe something, I would kill myself. It is one of the most boring things that you can imagine in your life. From my point of view, maybe it’s super fun, but from my point of view it is horrible. So, how could you get so many people transcribing?

David: It turns out that, first of all, many people already know how to do transcription in the world. A lot of them are professional transcribers or they’re paralegals, or folks who have a lot of experience listening to audio and typing really quickly and with high accuracy. At this point, we have such a brand awareness in the market that a lot of people come to us and want to do transcription work for us. Of course, we test them and evaluate their skill set and make sure that they can do the work.

But when we were starting out, what we did was we started on a platform called Amazon Mechanical Turk. Mechanical Turk is this marketplace from Amazon that is intended for doing the small units of work. We started putting transcription work out there and that helped us get our initial base of transcribers. Now of course, we have many people that come to us directly and want to do transcription work, but that’s how we got started.

Marco: How you kick it off. Another curiosity. Sorry, David, I’m very curious about a service that is very smooth from a customer point of view, but I mentioned is very complicated behind the scenes. I would like to understand it a little better how it works. What’s the percentage of the service that is human- powered, and what is the percentage that is automatic algorithm and so on?

David: All of the transcription itself is done by humans. So, you’re always getting a human being who is doing the transcription. Now, we have some other capabilities like time stamping where we insert time codes into the transcripts. We have capabilities for taking the text and outputting it into certain formats. These kinds of things are doing by the machine, if you will.

The actual transcription is done by the human, because humans are great at listening and recognizing audio and turning that into something written. People are really, really good at that. We make it easy for them to do that kind of work, and then we take care of all the other stuff, the conversion of the audio and video into the right format, the checking the rules. A bunch of the other things that are often time consuming. We do that with computers, but the high-quality transcription, that is always done by a person.

Marco: The person doing the transcription obviously gets a cut or gets rate compensation in a percentage way I imagine?

David: The way we do that is the transcribers are compensated for their work. We have an equation. We call it Transcription plus Review equals the final price we are paying. The higher quality the transcription is the less we spend on review. If there are some errors on some of the transcription, we’re spending more on the reviews. Those two things tend to balance each other out, so that we can ensure we are delivering a really, high-quality work product to the customer. That’s kind of how the system works.

So, if you’re a transcriber who is doing a lot of work, one time you might be doing a bunch of transcriptions and another time you are doing a bunch of reviews, but you are reviewing someone else’s work. The system has this really nice equilibrium and this nice market effect, this network effect where the more transcribers we get, the more customers we get. The more customers we get, we build up the transcriber base. We are always building up more and more customers and more and more transcribers, and that gives us this equilibrium, you know, this growth in the market as we call it.

Marco: Another thing that I was curious about, Dave, is the following: sometimes I shoot a video, and I need a very fast transcription. So I go on Speechpad and say that I need it in 24 hours. I’m ready to pay a higher price, but I need it fast, and I get it in 24 hours. I always think, “How the hell can they transcribe 40 minutes, maybe a one-hour video like this so fast and in such a correct way?” How can you handle this? I mean, with different clients, different languages, and different customers?

David: Just the way you think about the Amazon Cloud, let’s say. Amazon Web Services. Letting companies scale their compute requirements on demand. We provide a similar capability for scaling the human work force on demand. We are managing the work force, so we can sort of see, “Okay, here’s how much work there is.” We know there’s more and more. We know there is a spike, so we alert the transcribers that there is more work. They come in, take those jobs and get it done.

Now in the case of long-form video, like what you are talking about, we do sometimes break that into a couple of pieces. Say you have a 60-minute video. We might have two people work on that file at the same time. That way, you know, say in a 12-hour period, we are having two people work in parallel. In theory, someday, we could have 60 people work in parallel on one-minute pieces of audio and get a 60-minute transcription done in the time it takes one person to do one minute of audio. We could theoretically be doing an hour of audio or video in an hour.

Marco: What are the main reasons, in your opinion, David, for companies and for people doing business online to add transcriptions? Because a lot of colleagues or a lot of companies I’m talking with, they really underestimate, in my opinion, the power of transcription. My opinion is clear because I see for SEO reasons, for several reasons. What are your reasons, the most important reasons why transcription should be added in any kind of project?

David: Our biggest and fastest growing vertical is something called Video SEO. The advantage of doing that is that the text from that video is then indexed by Google and Bing and other search engines, and so that increases their rankings and drives more traffic to their websites. The great thing about video, like the video we are doing right now, is it’s very interactive. It’s very dynamic and people love to do it. People love video.

The challenge for the search engines is making this kind of content in a meaningful way. The search engines can’t find this video as easily. So when we provide the customer with the transcription and they put that on their site, it’s really easy for the search engines to index it.

Marco: How do you relate, for instance, YouTube offering now the automatic transcription now? Obviously, [Alda] is all automatic, I guess, is not human also because most of the time it’s totally wrong by the way. How do you relate with that?

David: We love it when people try other solutions like that because customers come back to us and say, “We tried the automated stuff or we tried another solution, and we need the quality that Speechpad delivers.” The real difference in what we provide is that we are always providing high-quality transcriptions, so what you hear and what people are saying, that is actually what you are getting. You are not getting a bunch of other words that people did not say. You are getting a very high-quality work product.

Marco: Tell me again Dave, how can you do the quality control? How can you be so focused on quality control? What do you do to [grant] it?

David: Though this is the way say, eBay has the rankings when you sell an item and you get feedback on that. Think of our transcription rating system the same way. Every time a transcriber does a transcription, we review that transcription, and they get a score, and we make it really easy for the transcriber to see any mistakes they made.

I think the first thing is that helps them improve over time. Secondly, we pay people based on the quality of their work. The better, the fewer the mistakes they make, the more they get paid. Third, this review process results in a score, and so over time, transcribers might be doing hundreds, thousands of transcriptions, and they are building up their score just like you do no eBay or other market places.

Marco: By the way, while you were speaking, I was thinking that an additional advantage and benefit I see in having a very good transcription is that when I upload the good transcription on YouTube, for instance, I get a better translation because then you can use the Google translation tool. If the transcription is horrible, then you have a horrible translation. This is also another interesting point.

Dave, what are your top advices to companies willing to have a transcription done? If any, I mean? What are your takes to help companies having good transcriptions out of their work?

Marco: I think one of the biggest things people can do is set up a regular schedule to do videos. It doesn’t have to be hours and hours of video. It could literally be-we work with a company called SEOmoz, which provides software for SEO. They do something called a Whiteboard Friday. It’s a relatively short piece of audio. They do it every Friday. They share tips and tricks with their customers. That’s a really compelling way for a company to create some content. It’s easy to create. It’s quick to create. Then, the transcription gets done quickly. You put that up on the web, and all of a sudden, you are getting a lot more pages indexed in Google or Bing. I think that’s one really great thing.

The other thing is customers may already have video assets that they can work with. A lot of people have recorded video about their product launch or an interview with their head of engineering or their CTO or a video with a customer they did in the past. That’s all great content. You don’t necessarily have to create new content. You can get that transcribed, then all of a sudden, you are getting a double value from that content you already built in the past.

There is a lot of content customers already have in the form of audio and video that you can get transcribed and start getting ranked for in the search engine.

Marco: This is cool because you can use also all your archives. I never thought about that, so it is very smart. By the way I had Rand Fishkin guest at Tech Alchemist and really appreciate his job.

All right. About format, no problem because you said that you transcribe any kind of format, so it’s no problem to worry about any particular format. Which language Speechpad can transcribe? Only English?

David: We are heavily focused on English. We are starting to do Spanish now. We have a lot of demand from customers to do Spanish. We certainly get a lot of inquiries for French, German, Japanese, Korean, lots of other languages. We are looking at adding those next year.

Marco: All right. I’m waiting for your Chinese version. That would be big fun.

David: We have a lot of customers who request that. What we tend to do, we do run some of that through the system, but to do it in really high volume, we’ll need to scale up the workforce also for those other languages.

Marco: I always thought-a few years ago I had the video translated in six to seven different languages. I couldn’t understand Mandarin, so I had no idea if it was good or bad. This is another…

David: Exactly. Yeah, a lot of people want to do the other languages for translation purposes, or they have video content, and they want to do subtitling, that’s a big category for us. Or closed captioning added to the video, so that is something we’ll spend more and more time on next year.

Marco: One thing I didn’t mention about Speechpad, which I think is very useful, is that when I download the transcription, I get several formats. So I can have it in [inaudible 21:55] I can have the HTML, RTF, so several formats, and it’s useful if I have to use it on different media for subtitling or in a doc Word or something like that.

Right. Dave, a few minutes more. I want to talk about big data with you too. You recently wrote a good post that I suggested to the Tech Alchemist community on Forbes, if I’m not wrong. You are a contributor with the blog on Forbes where you talk about several stuff.

Dave: Yeah, exactly.

Not only Speechpad, but you have a long career as a successful entrepreneur and writer. So guys just go there and watch all the Dave job because otherwise we stay here talking for hours.

How about big data? What’s happening right now?

David: Thanks a lot. If you think about all the-let’s take it in the context of Speechpad, you think about all the audio and video we work with everyday, thousands and thousands of files upload. You can imagine looking at the transcripts for keywords. You could do other analytics on all this audio and video and that can give you some insight.

Let’s take an example, you know, let’s say an insurance company might have tens of thousands of transcripts of recorded statements that they are taking from their clients over the course of a year. Imagine that they then analyze all those tens of thousands of transcripts looking for trends in the words and finding things like left turn is always associated with a certain issue that they know of in their system, a certain accident type or a certain kind of insurance claim or what have you. I’m just giving that as an example.

So that kind of analytics based on these huge data sets is a good example of what is going on in big data. That kind of shows the power of taking lots of data that might not have been in the right format to work with in the past. Audio in its native format is very hard to work with, but once you convert it into text, you can run all these incredible analytics on it to get these business insights that could reduce your business costs or drive more sales or what have you. That’s kind of the power of big data.

For my column, I tend to talk to lots of different startups and different companies in the space, and then I write about things that I see in the space and emerging trends.

Marco: By the way Dave, I always ask the Tech Alchemist guests to suggest their favorite two to three websites or apps that they can’t live without for business, of course. For productivity or for something that you think, “Gosh, I can’t play without that.” And for online digital business, “Absolutely, I recommend to use that kind of app.” Do you have your favorite two or three?

David: Well, I’m going to take a little different angle on that. I’m going to tell you my experience using a Mac. When I was 16, I went to work at a company called Microsoft. I really learned a ton, a ton of stuff.

Marco: I heard about that. Microsoft, I think. Yeah. I heard it.

David: Yeah. And listen, the company is huge and they have all kinds of productivity software. Windows 8 coming out, things like that. I love the recent work that they are doing, but I have to say a little while ago I was like, “I’m going to try and be a Mac user.” The most amazing thing to me is that my presentations look really great. People look at my slide decks, and they go, “Wow, great design work.” I never thought of myself as a designer. I think of myself as a product guy. I think of myself as a technologist. I occasionally do sales with some of our customers, things like that. But I never thought about being a great designer. One of the amazing things for me is I too can produce stuff that looks really good. That’s been really transformative for me. That’s one thing I can’t live without.

The other is not so much an app, but I’m an Ironman. I just did an event called Ironman France in June. If you think about people who are into sports and athletics, we are big data people. We love collecting data about ourselves, about calorie intake, our exercise habits, and how many calories we are burning, and how many feet we climb on the bike course, and all kinds of other data. The great thing for me is I can collect all that data using a couple hundred dollar GPS watch from Garmin, and I’m collecting immense amounts of data about my activities, and I’m uploading them into the Cloud, and I get to see everything about my activities. That’s an amazing thing to me that I can collect such granular data, and then view it and get insights from it. Those are really the two things I can’t live without.

Marco: This is, by the way, is very cool. This trend called Quantitative Self, so the possibility that you have to measure yourself with all the tools, FitBit and so on. The next web conference for people in Europe in December will be [inaudible 27:44] of things and all about this kind of stuff. It’s absolutely interesting.

Can you imagine David, that in the notes for the interview today about you, I was writing down avid triathlete and violinist. This was one thing. An inventor on 15 U.S. patents, that I thought, “Gosh, 15 U.S. patents is a lot, you know.”

David: It’s important to stress co-inventor because whenever you are inventing something it’s a lot of really, a team coming together to turn that invention into reality. It’s one thing to think about, “Hey, let’s build a transcription service that’s going to be really easy.” It’s another thing to make that a reality for a lot of customers. You got to have a lot of people coming together as a team to make that vision into reality.

Marco: Last question and then I’ll let you go. What will happen in the next couple of years in the transcription industry? I mean, these are exploding. I imagine the more videos there are, the more transcriptions there will be. There is any particular trend in this industry?

David: I think that’s right. I think that video is going to be a huge source of growth. First of all, because it’s a lot easier to capture video. We’ve all got a device like this device where it’s easy to capture video. That’s creating a huge growth in the amount of video.

Also, people are putting it online. That is one of the big differences I think. We take for granted that of course I can upload a video and put it on the Internet, but if you think about it, that is a relatively recent change which is not just I could capture it, but that I had enough bandwidth so that I could upload that and put it somewhere where people could get to it. That change is really disruptive in what’s going on. That combined with a bunch of regulatory stuff around requirements for closed captioning to make video accessible, and just the sheer volume of audio and video being put on the net, I think that’s where a lot of the growth is going to come from. As a company, our goal is continuing to scale and provide work for more and more people. There is a lot of unemployment in the country right now, and so we view one of the things that we are doing as creating a place where people can do productive work, get paid for it, have a career path potentially. So that’s where we are going to be all about for the next few years is continuing to scale the business while maintaining high quality.

Marco: Thank you so much, Dave. David Feinleib, CEO and Co-Founder of Speechpad and good luck for everything. Keep in touch.

David: Marco, thanks so much.