I have been enjoying Mark Frauenfelder’s transcription software links at BoingBoing over the past couple of days, and had reason to correspond with Mark about something unrelated this week. In the course of the correspondance, I mentioned to Mark how I was enjoying his stuff, and that he might be interested in my homebrewed transcription solution.

I also mentioned that on my last round of interviews, I had explored online transcription services; given the non-existent budgets I have access to, the lowest rates were the absolute determinant in my search. It was also one-hundred percent necessary that I be able to upload audio files directly to the service; if the account setup process was also fully automated, then I was in heaven.

In the end, I settled on both iDictate.com and escriptionsist.com. iDictate’s primary rate is one cent per word, and they offer same-day regular service turnaround (but don’t get too excited, because there are some caveats). Escriptionist offers a flat rate based on the length of the audio files; that rate is $50 per half hour of audio.

Both offer uploads.

iDictate’s fast turnaround and low rates are apparently enabled by offshoring; the scuttlebutt on the internets is that the service breaks up inbound audio files into shorter pieces and sends them to multiple typists. Whether this is true or not, the fast turnaround is quite factual. I submitted three files to iDictate.

All three files were two-speaker interviews conducted over the phone, and each file was approximately 30 minutes in length, in mp3 format. The first file I submitted was rejected for audio quality reasons; I was billed for the 400 words or so that had been transcribed and that appeared in my returned document. Checking my account, I saw that that failed experiment was going to cost me $4.

“Excellent!” I thought to myself, rubbing my hands together with glee. I decided to submit the rejected file to escriptionist.com, but would hold off until I had a completed file in hand from iDictate.

The next file I submitted was accepted, apparently – at any rate, I did not get a rapid rejection notice, accompanied by an incomplete transcription.

A day later, the file had been transcribed. It wasn’t pretty, and there were many instances of roughness in the file (skipped words denoted by asterisks, both speakers unidentified, etcectera). But it was more than sufficient for my needs. Excitedly, I submitted the next file, the second half of the interview. My eyes bugged out of my head when it returned to me in three hours.

I checked my account balance. It reported a total cost of about $100 for all three files. I fairly danced with glee.

Next, I initiated the submission process with escriptionist.com. The process was slow and unwieldy, requiring a personal phone call and many email messages before I was okayed to upload my 31mb mp3. For reasons unrelated to my current DSL problems, the upload took forever – the ftp server on their end was only accepting material at 6kb/sec. Realistically, this doesn’t matter if you can upload overnight. Psychologically, it was frustrating as all get out.

I settled down to wait. On the third day, I emailed to enquire if my file was done. It was, and was emailed promptly to me, along with a PDF invoice for $54, billed to my credit card.

On opening the file, I was overjoyed. While the iDictate files were quick-and-dirty, costing (I thought) a penny a word, and yielding about 5,000 words per half hour, the escriptionist file was meticulous, beautifully formatted, and scrupulously accurate. It was also 7000 words long, rendering the per word cost considerably less than 1 penny.

I thought I was done, and had happily established relationships with two differing, but comparable in value, services. iDictate’s speed makes it valuable in deadline-sensitive situations; escriptionist’s attention to detail makes it valuable if you have a week to wait.

Then I got my bill from iDictate. It was for over $200. I logged in and checked my account totals, which were slightly over $100. I fired off a note, requesting that the double billing be removed. I received a note in reply asking whether I hadn’t misunderstood the terms – the one-penny rate applies only to single-person, old-school, verbal composition. To dictation, in fact. The terms clearly state that two people having a conversation on the phone qualify for the two-penny rate.

I’d misunderstood the terms of service. I wrote back, accepting my mistake, and requested that my account be cancelled. Two cents can’t be justified under the rates I currently get, and I certainly don’t want the temptation hanging over my head the next time I’m procrastinating a transcription.

To my irritation, the correspondent wrote back with a cheery “That’s fine,” referring to my willingness to pay twice what I had expected to and what the site’s own publicly accessible billing tools reported regarding my balance. “There is no monthly fee,” he cheerily concluded, in what I took to be deliberate disregard of my instructions.

A few days later, another note arrived, telling me how to access my account while the service changes servers. I wrote back politely requesting that my cancellation request be honored.

So, in short, iDictate’s low base rate, fast turnaround and lack of competition mean that, for now, the service offers what I would charitably term sucky customer service. As someone helping to run a tiny business myself, I can’t say I don’t understand the business posture. But as a customer, I’m pretty pissed off, and won’t be using them again.

On the other hand, I was frustrated beyond comprehension during the setup and upload process associated with escriptionist.com; the three day wait was excruciating. But the material, when it arrived, was deeply satisfying, and very clearly a significantly better consumer value.

At any rate, I now no longer have the favorite excuse of the dawdling writer for turning an interview into an article – transcription is no fun at all, and highly time consuming (for me, four hours to get a half-hour interview is about right). Having access to the option of transcription is a giant psychological weight lifted from my shoulders as a writer.

Who knows? Maybe I’ll even try old-school verbal transcription at iDictate one day – on a new account, of course. If the total time to a first draft is cut by half, it’s clearly a possible route to compress composition time.

But on the whole, I’ll be carefully budgeting my time and fees to incorporate escriptionist into my copy development process. They delivered a better product for a better price.

14 thoughts on “Twenty-first Century Typist”

  1. Hi, Mike. Until a few months ago I transcribed many of the programs on IT Conversations. I was able to find a service in India that did the job for rates on the order of $50/hr of audio, or half of what Escriptionist charges. (Rates in India are based on 65-character lines, not words.) Because of the many technical terms, the transcriptions weren’t perfect. They always required another pass to make them accurate, usually with a handful of minor corrections per double-spaced page. I sent then MP3 files and a Word template I created, and they sent back the .doc file. …doug (producer, IT Conversations, http://www.itconversations.com)

  4. Great info…looking for any guidance on the first part of the equation – what do you use to capture the digital audio files? Maybe this is well known for media folks, but for anthro/user-research types, we actually don’t have a clue about it…this would be a fantastic way to go, but I don’t even know what the real tools are.

    Thanks.

  5. Steve:

    What Mark has been describing is (I think) capturing directly to his powerbook using one of the two digital apps he links to. I use SoundStudio to record audio in to my powerbook.

    I use the same telephone adapter Mark does to record phone interviews.

    I prefer to record the calls on a generic microcasette recorder than to record driectly to the powerbook, because the headphone jack on the casette recorder has a manual volume control, allowing greater control over levels when I record something into the computer.

  6. Is anyone worried about confidentiallity issues? Does the transcribing company sign an agreement regarding IP rights?

  7. > Is anyone worried about confidentiallity issues?
    > Does the transcribing company sign an agreement regarding IP rights?

    Frankly, this should be a non-issue. The person/company performing the task, to put it in a U.S.-centric fashion, is performing a “work for hire” and thus has no IP rights in the final product. For any company or individual performing transcription to violate confidentiality–or worse, claim IP rights in a document they have transcribed–would amount to nothing less than professional suicide.

    My policy is stricter than most–I don’t even reveal the identities of my customers, unless they specifically authorize it. Trust, service and confidentiality are the pillars upon which my business is built.

  8. I read with interest your article and the various posts. As the founder of A Word Away, Inc., a world-class provider of audio to text conversion, I have spent the last 15 years studying best practices, pricing methodologies and in general, all there is to know about transcription and captioning. Lest we forget, it’s all about exactly and methodically converting the spoken word (and sometimes accompanied by non-verbal components) into text — randomly referred to as transcription, closed captioning, court reporting and so many other garden variety terms.

    This industry, devoid of standards, was based since nascency on the basic sweatshop business model: a shrewd business person, usually male, at the helm, paying workers, usually female, at a fraction of the hourly fees.

    The usual suspects prevail over this industry and the best transcribers and captioners go where the money is — and the cream rises to the top.

    Confidentiality within this model is the least of one’s problems … one can imagine inner sanctum Wall Street discussions making it on to the front page of the New York Times but even more sinister, workers posing as transcriptionists methodically hijacking vital information and supplying it, at a price, to others.

    I’ve been very, very fortunate to perform services to an elite few who realize that there is no 98% quality, 95% quality — who are they kidding! — when it comes to a transcript.

    A transcript is either a) an exact, letter perfect, impeccable representation of what the speaker said; or, to use my famous coined phrase, b) pumping junk.

    The former will cost you plenty but will be securely (yes through a secure bulletproof VPN) delivered almost instantaneously, guaranteed faster than any competitor.

    The latter? I’ve seen dropped words, dropped paragraphs, sentences the speaker never said, alienated meanings … why would any consumer want to endure this? What’s the point?

    No one has mentioned the fact that you cannot “QC” a transcript. Most firms take rough draft output from low level transcribers and then try to run it through a proofreading stage … proofreading WHAT? You can’t apply proofreading to what should be a methodical conversion. You would have to go back against the recorded material to do that; that’s doing the transcript over again.

    Then of course there’s the hit or miss feature (will you get the same result each time you use etranscriptionist, idictate or the multitude of others).

    After many years of watching a displaced court reporting workforce and wondering why there had never been a bridge between transcription and court reporting, around six years ago I set out to create such an environment.

    I created a set of methodologies that finally cracks the code to a certain extent. All this just to allow me to add the multiplication factor, to duplicate what was basically a one person unique skill (mine) and transfer the skills and disciplines without dropping quality, performance or price performance.

    For now, I’ll happily remain the self-proclaimed “diva” of the audio to text conversion business and enjoy counseling others, remaining on the edge of best practices as they emerge…

    and from time to time, when certain customers demand (and they often do, saying, we don’t care so much about the quality, this is a huge, huge volume and we only want to run through it with a highlighter picking things up here and there) …

    I’ll even “pump junk” (hire faceless others to just produce something and charge the kinds of rates you described in your notes).

    Most satisfying for me, though, has been this certain faction of consumer who just realizes the cost/value equation and pays good and just rates with a smile.

    Honestly, good rates average around $250.00 per 60 mins of good (just plain good not perfect) audio, resulting in around 31 or 32 double-spaced pages of 11pt text, regardless of how many people are involved, with at least a two-day turnaround.

  9. I’ve been reading these posts with interest since it was our company that was mentioned in the article. I’m very appreciative for Mr. Whybark taking the time to evaluate our services and write about his findings. We’ve since upgraded our servers to avoid the hassle that he unfortunately went through and it’s this kind of honest feedback from our customers that we love. It only makes us better.

    In response to the post by Teresa, I’m delighted to let you know that our profit margins are thin, we pay our transcriptionists well (we have a waiting list of people wanting to work for us), and we operate on a general business model of integrity and honesty. We most recently had another writer review our services and Idictate.com and we won hands down.

    While the methodology described above I’m sure is precise, we stand by our word of 98% accuracy for reasonably good audio and it’s that accuracy along with our reasonable prices that keep all of our customers coming back again and again.

    What does make us different is that we are running a business, but aren’t interested in cranking out as many transcripts to make as much money as we can. It’s this approach to business that most of our customers have found refreshing (not to mention the quality of product being delivered). Plus, our clients can be assured that their audio is being transcribed by Americans who have a much better understanding of the English vernacular than most overseas outfits.

    I’m far from a shameless promoter and never post on these kinds of sites, but did want to put in my two cents worth in response to the above post.

    It sounds like you’re providing a very valuable service as well and we wish you continued success!

    J.R.

  10. I have agreed to transcribe about 12 hours of audio (I can’t remember at the moment, but I have it written down)…it took me about 50 hours of work, anyway. I am wondering what to charge…the man I am doing it for is a friend, but he will actually be paying me from money from a book publisher, so I want to give him a deal but not too good of a deal. The text is interviews between him and the writer of the book. Any help would be appreciated…I want to be able to give him what the average rate is, and then what my rate will be, to compare. Thanks!

  11. My firm uses CyberSecretaries –

    http://www.youdictate.com

    They charge a per word flat rate and only use typists from North America. I have never had any complaints and they go out of their way in terms of formatting.

  12. Hi,

    We are a provider of transcription services (Production Transcripts) for a wide range of audio types. We’re set up to efficently handle uploads and our servers can take it as fast as you can give it… 🙂

    Like others mentioned above, we price by the minute of audio and you can get quotes directly from our site. Our prices are as competetive as anybody else’s and our turnaround time is pretty darn quick for the price/accuracy.

    While I realize that a lot of focus now is on uploading files, many of the services mentioned fall flat if you actually have audio on a physical format. For us, it is no be deal and we can handle most audio and video formats without any delays in turnaround.

    We handle quite a few large clients and provide a fully functioning secure client area with access to your transcripts and invoices 24/7. That way when you’re on a deadline and need your transcript at four in the morning and don’t have the orignal that’s been e-mailed to you, you can just log in and download a fresh copy. This is particularly useful on large jobs with hundred of transcripts. It also allows you to set up additional users, so that everybody who need access to a particular job can access the material independently

  13. thank you for all this interesting info. i did not know, that there were any transcribing services, i had to do everything myself. well, i learned how to type when i was 15, it helps me still at 41. i keep my tapes and start to transcribe the moment i need the data. recently, it took me one day to write an interview of 90 minutes. i still work with my stereo and mp3 files on pc.

    i did not yet decide what to buy for speech recording. i was looking for a minidisc player, then i tried to use an mp3 player. well, they can record quite much, but not enough.

Comments are now closed.